Chapter 57

TimeGAN: Time-series Generative Adversarial Networks

NotebookPython 3 (ipykernel)49 cells

TimeGAN: Time-series Generative Adversarial Networks

Docker image: ml4t-gpu

Book Reference: Chapter 5, Section 5.4 (GANs for Financial Time Series)

GPU recommended: This notebook trains models with PyTorch/CUDA. It will run on CPU but training may be very slow. For GPU acceleration:
bash
docker compose run --rm ml4t-gpu python 05_synthetic_data/01_timegan.py

This notebook implements TimeGAN (Yoon, Jarrett & van der Schaar, NeurIPS 2019), the foundational architecture for synthetic financial time series generation.

Learning Objectives

Understand TimeGAN's five-component architecture and why latent-space training matters
Implement the three-phase training approach (embedding → supervisor → joint GAN)
Evaluate synthetic data using the Fidelity-Utility-Privacy framework
Apply Train-Synthetic-Test-Real (TSTR) validation with proper temporal splits

Why TimeGAN Matters

TimeGAN introduced two key innovations that address limitations of standard GANs:

Stepwise Supervised Loss: Standard GANs only learn the overall distribution. TimeGAN adds explicit supervision on temporal transitions (how t → t+1).
Learned Embedding Space: Instead of operating directly on raw data, TimeGAN learns a latent representation where adversarial training is more stable.

With 1,800+ citations, TimeGAN remains the baseline against which newer methods are compared.

Data Format

We use 6 stocks (BA, CAT, DIS, GE, IBM, KO) with adjusted close prices, matching the 2nd edition benchmark format. The multi-stock panel exposes the model to a wider distribution of volatility and trend regimes than single-stock OHLCV.

References

Paper: Yoon, J., Jarrett, D., & van der Schaar, M. (2019). "Time-series Generative Adversarial Networks." NeurIPS 2019.
Official Code: https://github.com/jsyoon0823/TimeGAN

In [1]python · cell 2

python

"""TimeGAN — Time-series Generative Adversarial Networks."""

import json
from datetime import UTC, datetime
from pathlib import Path

import joblib
import matplotlib.pyplot as plt
import numpy as np
import polars as pl
import torch
import torch.nn as nn
import torch.optim as optim
from IPython.display import Image, display
from sklearn.preprocessing import MinMaxScaler
from timegan_metrics import run_timegan_evaluation
from torch.utils.data import DataLoader, TensorDataset
from tqdm import tqdm

from data import load_us_equities
from utils.paths import get_chapter_dir, get_output_dir
from utils.reproducibility import set_global_seeds
from utils.style import COLORS, plot_fidelity_comparison

In [2]python · cell 3

python

TRAIN_STEPS = 10000  # Steps per training phase (matching official repo)
RETRAIN = True  # Set True to force re-training even if checkpoint exists
SEED = 42

In [3]python · cell 4

python

set_global_seeds(SEED)

# Paths
ASSETS_DIR = get_chapter_dir(5) / "assets"
OUTPUT_DIR = get_output_dir(5, "timegan")
CHECKPOINT_DIR = OUTPUT_DIR / "checkpoints" / "timegan" / "multi_stock"

In [4]python · cell 5

python

# Configuration (matching 2nd edition benchmark)
#
# Data choice rationale: the six-stock panel spans different volatility/trend
# regimes, which gives the embedder a wider feature distribution than the four
# OHLCV columns of a single stock would.
TICKERS = ["BA", "CAT", "DIS", "GE", "IBM", "KO"]
SEQ_LEN = 24
HIDDEN_DIM = 24
NUM_LAYERS = 3
BATCH_SIZE = 128
LEARNING_RATE = 1e-3

device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
print(f"Using device: {device}")

Output

Using device: cuda

TimeGAN Architecture

TimeGAN consists of five interconnected modules that operate in a shared latent space:

Module	Purpose	Training Phase
Embedder	Maps raw data → latent space	Phase 1 (autoencoder)
Recovery	Reconstructs latent → raw data	Phase 1 (autoencoder)
Supervisor	Predicts next latent step	Phase 2 (temporal)
Generator	Produces latent sequences from noise	Phase 3 (joint GAN)
Discriminator	Classifies real vs fake latent sequences	Phase 3 (joint GAN)

In [5]python · cell 7

python

if (ASSETS_DIR / "timegan_architecture.jpeg").exists():
    display(Image(ASSETS_DIR / "timegan_architecture.jpeg", width=700))

1. Load Data

We use 6 diverse stocks with adjusted close prices. This provides:

Diverse dynamics: Each stock has different volatility and trends
Consistent scale: All normalized to [0, 1]
Learnable patterns: Cross-asset relationships are meaningful

In [6]python · cell 9

python

def load_multi_stock_data(
    tickers: list[str], start_year: str = "2000"
) -> tuple[np.ndarray, np.ndarray]:
    """Load adjusted close prices for multiple stocks."""
    df_pl = load_us_equities()
    df_pl = df_pl.filter(pl.col("symbol").is_in(tickers))

    # Update tickers to only those actually present in the data
    available = df_pl["symbol"].unique().to_list()
    tickers = [t for t in tickers if t in available]
    if not tickers:
        raise ValueError(
            f"None of the requested tickers found in data. Available: {available[:10]}"
        )

    # Pivot to wide format
    df = (
        df_pl.select(["timestamp", "symbol", "adj_close"])
        .pivot(on="symbol", index="timestamp", values="adj_close")
        .sort("timestamp")
        .to_pandas()
        .set_index("timestamp")
        .loc[start_year:, tickers]  # Ensure column order matches tickers
        .dropna()
    )

    timestamps = df.index.to_numpy()
    data = df.values.astype(np.float32)

    print(f"Loaded {len(df)} rows, {len(tickers)} stocks")
    print(f"Date range: {timestamps[0]} to {timestamps[-1]}")
    print(f"Stocks: {', '.join(tickers)}")

    return data, timestamps

In [7]python · cell 10

python

all_data, all_timestamps = load_multi_stock_data(TICKERS)
n_features = all_data.shape[1]  # Actual number of stocks found (may differ from TICKERS)
print(f"Shape: {all_data.shape} (days × stocks)")

Output

Loaded 4586 rows, 6 stocks
Date range: 2000-01-03T00:00:00.000 to 2018-03-27T00:00:00.000
Stocks: BA, CAT, DIS, GE, IBM, KO
Shape: (4586, 6) (days × stocks)

Temporal Train/Holdout Split

We split temporally to enable unbiased TSTR evaluation. The generator never sees holdout data during training.

In [8]python · cell 12

python

# Use last 20% as holdout
n_train = int(len(all_data) * 0.8)
train_data = all_data[:n_train]
holdout_data = all_data[n_train:]
train_timestamps = all_timestamps[:n_train]
holdout_timestamps = all_timestamps[n_train:]

print(f"Training:  {len(train_data):,} days ({train_timestamps[0]} to {train_timestamps[-1]})")
print(
    f"Holdout:   {len(holdout_data):,} days ({holdout_timestamps[0]} to {holdout_timestamps[-1]})"
)

Output

Training:  3,668 days (2000-01-03T00:00:00.000 to 2014-08-01T00:00:00.000)
Holdout:   918 days (2014-08-04T00:00:00.000 to 2018-03-27T00:00:00.000)

2. Normalize and Create Sequences

In [9]python · cell 14

python

# Normalize to [0, 1] using MinMaxScaler (matching 2nd edition)
scaler = MinMaxScaler()
train_scaled = scaler.fit_transform(train_data).astype(np.float32)
holdout_scaled = scaler.transform(holdout_data).astype(np.float32)

print(f"Scaled data range: [{train_scaled.min():.4f}, {train_scaled.max():.4f}]")

Output

Scaled data range: [0.0000, 1.0000]

In [10]python · cell 15

python

def create_sequences(data: np.ndarray, seq_length: int) -> np.ndarray:
    """Create overlapping sequences from time series data."""
    sequences = []
    for i in range(len(data) - seq_length):
        sequences.append(data[i : i + seq_length])
    return np.array(sequences, dtype=np.float32)

In [11]python · cell 16

python

sequences = create_sequences(train_scaled, SEQ_LEN)
holdout_sequences = create_sequences(holdout_scaled, SEQ_LEN)

print(f"Created {len(sequences)} training sequences of length {SEQ_LEN}")
print(f"Created {len(holdout_sequences)} holdout sequences for TSTR")

# DataLoader
dataset = TensorDataset(torch.from_numpy(sequences))
dataloader = DataLoader(dataset, batch_size=BATCH_SIZE, shuffle=True)

Output

Created 3644 training sequences of length 24
Created 894 holdout sequences for TSTR

3. Model Components

We use ModuleList with separate GRU layers to match TensorFlow's layer-by-layer construction exactly. This ensures consistent behavior with the official implementation.

In [12]python · cell 18

python

class Embedder(nn.Module):
    """Maps raw sequences to latent space. Stacked GRU layers + Dense with sigmoid."""

    def __init__(self, input_dim: int, hidden_dim: int, num_layers: int):
        super().__init__()
        # Stack GRU layers manually to match TF behavior exactly
        self.gru_layers = nn.ModuleList()
        for i in range(num_layers):
            in_dim = input_dim if i == 0 else hidden_dim
            self.gru_layers.append(nn.GRU(in_dim, hidden_dim, batch_first=True))
        self.fc = nn.Linear(hidden_dim, hidden_dim)

    def forward(self, x: torch.Tensor) -> torch.Tensor:
        h = x
        for gru in self.gru_layers:
            h, _ = gru(h)
        return torch.sigmoid(self.fc(h))

In [13]python · cell 19

python

class Recovery(nn.Module):
    """Reconstructs sequences from latent space."""

    def __init__(self, hidden_dim: int, output_dim: int, num_layers: int):
        super().__init__()
        self.gru_layers = nn.ModuleList()
        for _ in range(num_layers):
            self.gru_layers.append(nn.GRU(hidden_dim, hidden_dim, batch_first=True))
        self.fc = nn.Linear(hidden_dim, output_dim)

    def forward(self, h: torch.Tensor) -> torch.Tensor:
        x = h
        for gru in self.gru_layers:
            x, _ = gru(x)
        return torch.sigmoid(self.fc(x))

In [14]python · cell 20

python

class Supervisor(nn.Module):
    """Predicts next latent step. Uses num_layers-1 layers."""

    def __init__(self, hidden_dim: int, num_layers: int):
        super().__init__()
        supervisor_layers = max(1, num_layers - 1)
        self.gru_layers = nn.ModuleList()
        for _ in range(supervisor_layers):
            self.gru_layers.append(nn.GRU(hidden_dim, hidden_dim, batch_first=True))
        self.fc = nn.Linear(hidden_dim, hidden_dim)

    def forward(self, h: torch.Tensor) -> torch.Tensor:
        s = h
        for gru in self.gru_layers:
            s, _ = gru(s)
        return torch.sigmoid(self.fc(s))

In [15]python · cell 21

python

class Generator(nn.Module):
    """Generates latent sequences from noise."""

    def __init__(self, input_dim: int, hidden_dim: int, num_layers: int):
        super().__init__()
        self.gru_layers = nn.ModuleList()
        for i in range(num_layers):
            in_dim = input_dim if i == 0 else hidden_dim
            self.gru_layers.append(nn.GRU(in_dim, hidden_dim, batch_first=True))
        self.fc = nn.Linear(hidden_dim, hidden_dim)

    def forward(self, z: torch.Tensor) -> torch.Tensor:
        e = z
        for gru in self.gru_layers:
            e, _ = gru(e)
        return torch.sigmoid(self.fc(e))

In [16]python · cell 22

python

class Discriminator(nn.Module):
    """Classifies real vs fake latent sequences."""

    def __init__(self, hidden_dim: int, num_layers: int):
        super().__init__()
        self.gru_layers = nn.ModuleList()
        for _ in range(num_layers):
            self.gru_layers.append(nn.GRU(hidden_dim, hidden_dim, batch_first=True))
        self.fc = nn.Linear(hidden_dim, 1)

    def forward(self, h: torch.Tensor) -> torch.Tensor:
        y = h
        for gru in self.gru_layers:
            y, _ = gru(y)
        return torch.sigmoid(self.fc(y))

4. Initialize Models

In [17]python · cell 24

python

embedder = Embedder(n_features, HIDDEN_DIM, NUM_LAYERS).to(device)
recovery = Recovery(HIDDEN_DIM, n_features, NUM_LAYERS).to(device)
supervisor = Supervisor(HIDDEN_DIM, NUM_LAYERS).to(device)
generator = Generator(n_features, HIDDEN_DIM, NUM_LAYERS).to(device)
discriminator = Discriminator(HIDDEN_DIM, NUM_LAYERS).to(device)


def count_params(model):
    return sum(p.numel() for p in model.parameters() if p.requires_grad)


print("Model Parameters:")
print(f"  Embedder:      {count_params(embedder):,}")
print(f"  Recovery:      {count_params(recovery):,}")
print(f"  Supervisor:    {count_params(supervisor):,}")
print(f"  Generator:     {count_params(generator):,}")
print(f"  Discriminator: {count_params(discriminator):,}")
total_params = sum(
    count_params(m) for m in [embedder, recovery, supervisor, generator, discriminator]
)
print(f"  Total:         {total_params:,}")

Output

Model Parameters:
  Embedder:      10,104
  Recovery:      10,950
  Supervisor:    7,800
  Generator:     10,104
  Discriminator: 10,825
  Total:         49,783

In [18]python · cell 25

python

# Check for existing checkpoint
CHECKPOINT_PATH = CHECKPOINT_DIR / "checkpoint.pt"
SKIP_TRAINING = False

if CHECKPOINT_PATH.exists() and not RETRAIN:
    print(f"Loading checkpoint from {CHECKPOINT_PATH}")
    checkpoint = torch.load(CHECKPOINT_PATH, map_location=device, weights_only=False)
    embedder.load_state_dict(checkpoint["embedder"])
    recovery.load_state_dict(checkpoint["recovery"])
    supervisor.load_state_dict(checkpoint["supervisor"])
    generator.load_state_dict(checkpoint["generator"])
    discriminator.load_state_dict(checkpoint["discriminator"])
    print("Checkpoint loaded - skipping training")
    SKIP_TRAINING = True

Three-Phase Training

TimeGAN uses a sequential training approach with step-based iteration (matching the official implementation):

Embedding Phase: Train Embedder + Recovery as autoencoder
Supervisor Phase: Train Supervisor to predict next latent step
Joint Phase: Train Generator + Discriminator adversarially

In [19]python · cell 27

python

if (ASSETS_DIR / "timegan_training.jpeg").exists():
    display(Image(ASSETS_DIR / "timegan_training.jpeg", width=800))

5. Phase 1: Embedding Training

Train Embedder + Recovery as autoencoder with loss: $10 \cdot \sqrt{\text{MSE}}$

In [20]python · cell 29

python

if not SKIP_TRAINING:
    print("\n" + "=" * 60)
    print("PHASE 1: Autoencoder Training")
    print("=" * 60)

    opt_autoencoder = optim.Adam(
        list(embedder.parameters()) + list(recovery.parameters()),
        lr=LEARNING_RATE,
    )
    mse_loss = nn.MSELoss()

    # Infinite iterator for step-based training
    def infinite_dataloader():
        while True:
            yield from dataloader

    data_iter = infinite_dataloader()
    embedding_losses = []

    for step in tqdm(range(TRAIN_STEPS), desc="Phase 1"):
        (batch,) = next(data_iter)
        batch = batch.to(device)

        h = embedder(batch)
        x_tilde = recovery(h)

        # Loss: 10 * sqrt(MSE) - matching 2nd edition exactly
        embedding_loss = mse_loss(x_tilde, batch)
        e_loss = 10.0 * torch.sqrt(embedding_loss)

        opt_autoencoder.zero_grad()
        e_loss.backward()
        opt_autoencoder.step()

        if step % 1000 == 0:
            embedding_losses.append(torch.sqrt(embedding_loss).item())
            print(f"  Step {step}: loss = {embedding_losses[-1]:.6f}", flush=True)

    print(f"Phase 1 complete. Final loss: {embedding_losses[-1]:.6f}")

Output


============================================================
PHASE 1: Autoencoder Training
============================================================

Phase 1:   0%|          | 0/10000 [00:00<?, ?it/s]

  Step 0: loss = 0.286357

Phase 1:   0%|          | 1/10000 [00:00<42:53,  3.88it/s]

Phase 1:   0%|          | 41/10000 [00:00<01:08, 144.52it/s]

Phase 1:   1%|          | 82/10000 [00:00<00:42, 233.46it/s]

Phase 1:   1%|          | 114/10000 [00:00<00:39, 250.15it/s]

Phase 1:   1%|▏         | 145/10000 [00:00<00:37, 259.41it/s]

Phase 1:   2%|▏         | 183/10000 [00:00<00:33, 294.52it/s]

Phase 1:   2%|▏         | 216/10000 [00:00<00:32, 301.44it/s]

Phase 1:   2%|▏         | 248/10000 [00:01<00:36, 265.23it/s]

Phase 1:   3%|▎         | 277/10000 [00:01<00:38, 255.72it/s]

Phase 1:   3%|▎         | 304/10000 [00:01<00:37, 258.96it/s]

Phase 1:   3%|▎         | 341/10000 [00:01<00:33, 288.27it/s]

Phase 1:   4%|▍         | 375/10000 [00:01<00:31, 301.85it/s]

Phase 1:   4%|▍         | 406/10000 [00:01<00:32, 295.78it/s]

Phase 1:   4%|▍         | 442/10000 [00:01<00:30, 313.00it/s]

Phase 1:   5%|▍         | 485/10000 [00:01<00:27, 343.43it/s]

Phase 1:   5%|▌         | 520/10000 [00:01<00:28, 337.70it/s]

Phase 1:   6%|▌         | 555/10000 [00:02<00:31, 298.63it/s]

Phase 1:   6%|▌         | 586/10000 [00:02<00:35, 267.46it/s]

Phase 1:   6%|▋         | 626/10000 [00:02<00:31, 300.18it/s]

Phase 1:   7%|▋         | 668/10000 [00:02<00:28, 331.64it/s]

Phase 1:   7%|▋         | 711/10000 [00:02<00:26, 357.11it/s]

Phase 1:   8%|▊         | 754/10000 [00:02<00:24, 375.45it/s]

Phase 1:   8%|▊         | 795/10000 [00:02<00:23, 383.57it/s]

Phase 1:   8%|▊         | 838/10000 [00:02<00:23, 394.72it/s]

Phase 1:   9%|▉         | 880/10000 [00:02<00:22, 401.87it/s]

Phase 1:   9%|▉         | 922/10000 [00:02<00:22, 404.66it/s]

Phase 1:  10%|▉         | 963/10000 [00:03<00:22, 400.94it/s]

  Step 1000: loss = 0.057189

Phase 1:  10%|█         | 1004/10000 [00:03<00:22, 401.24it/s]

Phase 1:  10%|█         | 1047/10000 [00:03<00:21, 407.30it/s]

Phase 1:  11%|█         | 1089/10000 [00:03<00:21, 410.50it/s]

Phase 1:  11%|█▏        | 1132/10000 [00:03<00:21, 413.91it/s]

Phase 1:  12%|█▏        | 1174/10000 [00:03<00:21, 414.99it/s]

Phase 1:  12%|█▏        | 1216/10000 [00:03<00:21, 400.72it/s]

Phase 1:  13%|█▎        | 1257/10000 [00:03<00:21, 399.35it/s]

Phase 1:  13%|█▎        | 1298/10000 [00:03<00:21, 396.24it/s]

Phase 1:  13%|█▎        | 1341/10000 [00:04<00:21, 403.97it/s]

Phase 1:  14%|█▍        | 1383/10000 [00:04<00:21, 407.28it/s]

Phase 1:  14%|█▍        | 1426/10000 [00:04<00:20, 411.54it/s]

Phase 1:  15%|█▍        | 1469/10000 [00:04<00:20, 414.84it/s]

Phase 1:  15%|█▌        | 1512/10000 [00:04<00:20, 417.18it/s]

Phase 1:  16%|█▌        | 1555/10000 [00:04<00:20, 418.74it/s]

Phase 1:  16%|█▌        | 1598/10000 [00:04<00:20, 419.25it/s]

Phase 1:  16%|█▋        | 1641/10000 [00:04<00:19, 420.39it/s]

Phase 1:  17%|█▋        | 1684/10000 [00:04<00:19, 420.55it/s]

Phase 1:  17%|█▋        | 1727/10000 [00:04<00:19, 421.15it/s]

Phase 1:  18%|█▊        | 1770/10000 [00:05<00:19, 421.26it/s]

Phase 1:  18%|█▊        | 1813/10000 [00:05<00:19, 421.31it/s]

Phase 1:  19%|█▊        | 1856/10000 [00:05<00:19, 421.73it/s]

Phase 1:  19%|█▉        | 1899/10000 [00:05<00:19, 421.77it/s]

Phase 1:  19%|█▉        | 1942/10000 [00:05<00:19, 421.44it/s]

Phase 1:  20%|█▉        | 1985/10000 [00:05<00:19, 421.52it/s]

  Step 2000: loss = 0.046126

Phase 1:  20%|██        | 2028/10000 [00:05<00:18, 420.85it/s]

Phase 1:  21%|██        | 2071/10000 [00:05<00:18, 420.71it/s]

Phase 1:  21%|██        | 2114/10000 [00:05<00:19, 413.02it/s]

Phase 1:  22%|██▏       | 2156/10000 [00:05<00:19, 406.21it/s]

Phase 1:  22%|██▏       | 2198/10000 [00:06<00:19, 409.14it/s]

Phase 1:  22%|██▏       | 2241/10000 [00:06<00:18, 412.37it/s]

Phase 1:  23%|██▎       | 2284/10000 [00:06<00:18, 415.04it/s]

Phase 1:  23%|██▎       | 2326/10000 [00:06<00:18, 416.38it/s]

Phase 1:  24%|██▎       | 2369/10000 [00:06<00:18, 417.88it/s]

Phase 1:  24%|██▍       | 2411/10000 [00:06<00:18, 418.24it/s]

Phase 1:  25%|██▍       | 2453/10000 [00:06<00:18, 418.42it/s]

Phase 1:  25%|██▍       | 2495/10000 [00:06<00:17, 418.40it/s]

Phase 1:  25%|██▌       | 2538/10000 [00:06<00:17, 419.18it/s]

Phase 1:  26%|██▌       | 2581/10000 [00:06<00:17, 419.82it/s]

Phase 1:  26%|██▌       | 2623/10000 [00:07<00:17, 418.79it/s]

Phase 1:  27%|██▋       | 2666/10000 [00:07<00:17, 421.55it/s]

Phase 1:  27%|██▋       | 2709/10000 [00:07<00:18, 402.83it/s]

Phase 1:  28%|██▊       | 2750/10000 [00:07<00:17, 403.44it/s]

Phase 1:  28%|██▊       | 2792/10000 [00:07<00:17, 406.48it/s]

Phase 1:  28%|██▊       | 2833/10000 [00:07<00:17, 404.04it/s]

Phase 1:  29%|██▉       | 2875/10000 [00:07<00:17, 408.56it/s]

Phase 1:  29%|██▉       | 2918/10000 [00:07<00:17, 412.21it/s]

Phase 1:  30%|██▉       | 2960/10000 [00:07<00:16, 414.51it/s]

  Step 3000: loss = 0.037880

Phase 1:  30%|███       | 3002/10000 [00:08<00:16, 415.92it/s]

Phase 1:  30%|███       | 3045/10000 [00:08<00:16, 418.12it/s]

Phase 1:  31%|███       | 3087/10000 [00:08<00:16, 418.39it/s]

Phase 1:  31%|███▏      | 3130/10000 [00:08<00:16, 419.04it/s]

Phase 1:  32%|███▏      | 3173/10000 [00:08<00:16, 419.38it/s]

Phase 1:  32%|███▏      | 3216/10000 [00:08<00:16, 420.07it/s]

Phase 1:  33%|███▎      | 3259/10000 [00:08<00:16, 419.29it/s]

Phase 1:  33%|███▎      | 3301/10000 [00:08<00:17, 382.32it/s]

Phase 1:  33%|███▎      | 3340/10000 [00:08<00:21, 304.21it/s]

Phase 1:  34%|███▍      | 3378/10000 [00:09<00:20, 320.56it/s]

Phase 1:  34%|███▍      | 3418/10000 [00:09<00:19, 338.80it/s]

Phase 1:  35%|███▍      | 3461/10000 [00:09<00:18, 361.16it/s]

Phase 1:  35%|███▍      | 3499/10000 [00:09<00:19, 341.03it/s]

Phase 1:  35%|███▌      | 3535/10000 [00:09<00:18, 340.28it/s]

Phase 1:  36%|███▌      | 3574/10000 [00:09<00:18, 352.50it/s]

Phase 1:  36%|███▌      | 3616/10000 [00:09<00:17, 369.33it/s]

Phase 1:  37%|███▋      | 3656/10000 [00:09<00:16, 376.36it/s]

Phase 1:  37%|███▋      | 3695/10000 [00:09<00:18, 343.29it/s]

Phase 1:  37%|███▋      | 3735/10000 [00:10<00:17, 357.30it/s]

Phase 1:  38%|███▊      | 3775/10000 [00:10<00:16, 369.06it/s]

Phase 1:  38%|███▊      | 3815/10000 [00:10<00:16, 377.02it/s]

Phase 1:  39%|███▊      | 3858/10000 [00:10<00:15, 391.03it/s]

Phase 1:  39%|███▉      | 3898/10000 [00:10<00:15, 392.37it/s]

Phase 1:  39%|███▉      | 3938/10000 [00:10<00:15, 392.13it/s]

Phase 1:  40%|███▉      | 3978/10000 [00:10<00:15, 391.61it/s]

  Step 4000: loss = 0.030111

Phase 1:  40%|████      | 4019/10000 [00:10<00:15, 394.31it/s]

Phase 1:  41%|████      | 4059/10000 [00:10<00:15, 384.58it/s]

Phase 1:  41%|████      | 4099/10000 [00:10<00:15, 386.72it/s]

Phase 1:  41%|████▏     | 4140/10000 [00:11<00:14, 392.95it/s]

Phase 1:  42%|████▏     | 4181/10000 [00:11<00:14, 397.37it/s]

Phase 1:  42%|████▏     | 4222/10000 [00:11<00:14, 399.40it/s]

Phase 1:  43%|████▎     | 4262/10000 [00:11<00:14, 396.99it/s]

Phase 1:  43%|████▎     | 4302/10000 [00:11<00:14, 381.15it/s]

Phase 1:  43%|████▎     | 4344/10000 [00:11<00:14, 391.23it/s]

Phase 1:  44%|████▍     | 4385/10000 [00:11<00:14, 393.99it/s]

Phase 1:  44%|████▍     | 4425/10000 [00:11<00:14, 395.53it/s]

Phase 1:  45%|████▍     | 4465/10000 [00:11<00:14, 387.79it/s]

Phase 1:  45%|████▌     | 4504/10000 [00:12<00:15, 351.97it/s]

Phase 1:  45%|████▌     | 4540/10000 [00:12<00:18, 302.51it/s]

Phase 1:  46%|████▌     | 4574/10000 [00:12<00:17, 310.62it/s]

Phase 1:  46%|████▌     | 4607/10000 [00:12<00:17, 307.30it/s]

Phase 1:  46%|████▋     | 4639/10000 [00:12<00:17, 308.78it/s]

Phase 1:  47%|████▋     | 4674/10000 [00:12<00:16, 318.61it/s]

Phase 1:  47%|████▋     | 4713/10000 [00:12<00:15, 336.90it/s]

Phase 1:  48%|████▊     | 4754/10000 [00:12<00:14, 357.44it/s]

Phase 1:  48%|████▊     | 4794/10000 [00:12<00:14, 369.58it/s]

Phase 1:  48%|████▊     | 4836/10000 [00:13<00:13, 383.38it/s]

Phase 1:  49%|████▉     | 4877/10000 [00:13<00:13, 388.34it/s]

Phase 1:  49%|████▉     | 4917/10000 [00:13<00:14, 339.66it/s]

Phase 1:  50%|████▉     | 4953/10000 [00:13<00:14, 342.42it/s]

Phase 1:  50%|████▉     | 4991/10000 [00:13<00:14, 351.82it/s]

  Step 5000: loss = 0.026226

Phase 1:  50%|█████     | 5029/10000 [00:13<00:13, 359.15it/s]

Phase 1:  51%|█████     | 5071/10000 [00:13<00:13, 376.42it/s]

Phase 1:  51%|█████     | 5114/10000 [00:13<00:12, 390.03it/s]

Phase 1:  52%|█████▏    | 5154/10000 [00:13<00:12, 392.84it/s]

Phase 1:  52%|█████▏    | 5196/10000 [00:13<00:12, 399.45it/s]

Phase 1:  52%|█████▏    | 5239/10000 [00:14<00:11, 406.16it/s]

Phase 1:  53%|█████▎    | 5280/10000 [00:14<00:11, 404.11it/s]

Phase 1:  53%|█████▎    | 5321/10000 [00:14<00:11, 394.07it/s]

Phase 1:  54%|█████▎    | 5362/10000 [00:14<00:11, 396.41it/s]

Phase 1:  54%|█████▍    | 5404/10000 [00:14<00:11, 401.53it/s]

Phase 1:  54%|█████▍    | 5446/10000 [00:14<00:11, 405.84it/s]

Phase 1:  55%|█████▍    | 5487/10000 [00:14<00:11, 406.64it/s]

Phase 1:  55%|█████▌    | 5528/10000 [00:14<00:11, 401.74it/s]

Phase 1:  56%|█████▌    | 5569/10000 [00:14<00:11, 402.77it/s]

Phase 1:  56%|█████▌    | 5610/10000 [00:15<00:10, 399.77it/s]

Phase 1:  57%|█████▋    | 5651/10000 [00:15<00:10, 395.88it/s]

Phase 1:  57%|█████▋    | 5693/10000 [00:15<00:10, 401.48it/s]

Phase 1:  57%|█████▋    | 5734/10000 [00:15<00:10, 398.21it/s]

Phase 1:  58%|█████▊    | 5777/10000 [00:15<00:10, 407.00it/s]

Phase 1:  58%|█████▊    | 5820/10000 [00:15<00:10, 411.74it/s]

Phase 1:  59%|█████▊    | 5862/10000 [00:15<00:10, 409.53it/s]

Phase 1:  59%|█████▉    | 5903/10000 [00:15<00:10, 406.65it/s]

Phase 1:  59%|█████▉    | 5945/10000 [00:15<00:09, 408.35it/s]

Phase 1:  60%|█████▉    | 5986/10000 [00:15<00:09, 403.76it/s]

  Step 6000: loss = 0.023576

Phase 1:  60%|██████    | 6027/10000 [00:16<00:09, 398.67it/s]

Phase 1:  61%|██████    | 6069/10000 [00:16<00:09, 404.15it/s]

Phase 1:  61%|██████    | 6111/10000 [00:16<00:09, 408.65it/s]

Phase 1:  62%|██████▏   | 6154/10000 [00:16<00:09, 412.05it/s]

Phase 1:  62%|██████▏   | 6196/10000 [00:16<00:09, 414.28it/s]

Phase 1:  62%|██████▏   | 6238/10000 [00:16<00:09, 415.41it/s]

Phase 1:  63%|██████▎   | 6280/10000 [00:16<00:08, 416.59it/s]

Phase 1:  63%|██████▎   | 6322/10000 [00:16<00:08, 411.70it/s]

Phase 1:  64%|██████▎   | 6364/10000 [00:16<00:08, 409.72it/s]

Phase 1:  64%|██████▍   | 6405/10000 [00:16<00:09, 396.29it/s]

Phase 1:  64%|██████▍   | 6446/10000 [00:17<00:08, 398.86it/s]

Phase 1:  65%|██████▍   | 6486/10000 [00:17<00:09, 389.22it/s]

Phase 1:  65%|██████▌   | 6527/10000 [00:17<00:08, 394.41it/s]

Phase 1:  66%|██████▌   | 6569/10000 [00:17<00:08, 400.80it/s]

Phase 1:  66%|██████▌   | 6612/10000 [00:17<00:08, 406.78it/s]

Phase 1:  67%|██████▋   | 6654/10000 [00:17<00:08, 410.33it/s]

Phase 1:  67%|██████▋   | 6697/10000 [00:17<00:07, 413.50it/s]

Phase 1:  67%|██████▋   | 6739/10000 [00:17<00:07, 415.01it/s]

Phase 1:  68%|██████▊   | 6781/10000 [00:17<00:07, 410.54it/s]

Phase 1:  68%|██████▊   | 6823/10000 [00:17<00:07, 412.61it/s]

Phase 1:  69%|██████▊   | 6866/10000 [00:18<00:07, 414.83it/s]

Phase 1:  69%|██████▉   | 6908/10000 [00:18<00:07, 416.13it/s]

Phase 1:  70%|██████▉   | 6950/10000 [00:18<00:07, 416.80it/s]

Phase 1:  70%|██████▉   | 6993/10000 [00:18<00:07, 417.78it/s]

  Step 7000: loss = 0.021318

Phase 1:  70%|███████   | 7035/10000 [00:18<00:09, 319.43it/s]

Phase 1:  71%|███████   | 7071/10000 [00:18<00:09, 311.55it/s]

Phase 1:  71%|███████   | 7112/10000 [00:18<00:08, 335.86it/s]

Phase 1:  72%|███████▏  | 7152/10000 [00:18<00:08, 351.50it/s]

Phase 1:  72%|███████▏  | 7194/10000 [00:19<00:07, 368.25it/s]

Phase 1:  72%|███████▏  | 7236/10000 [00:19<00:07, 381.78it/s]

Phase 1:  73%|███████▎  | 7278/10000 [00:19<00:06, 392.45it/s]

Phase 1:  73%|███████▎  | 7320/10000 [00:19<00:06, 399.46it/s]

Phase 1:  74%|███████▎  | 7362/10000 [00:19<00:06, 405.13it/s]

Phase 1:  74%|███████▍  | 7403/10000 [00:19<00:06, 402.40it/s]

Phase 1:  74%|███████▍  | 7444/10000 [00:19<00:06, 397.05it/s]

Phase 1:  75%|███████▍  | 7485/10000 [00:19<00:06, 398.94it/s]

Phase 1:  75%|███████▌  | 7529/10000 [00:19<00:06, 408.63it/s]

Phase 1:  76%|███████▌  | 7572/10000 [00:19<00:05, 412.36it/s]

Phase 1:  76%|███████▌  | 7614/10000 [00:20<00:05, 413.60it/s]

Phase 1:  77%|███████▋  | 7656/10000 [00:20<00:05, 411.43it/s]

Phase 1:  77%|███████▋  | 7698/10000 [00:20<00:05, 408.46it/s]

Phase 1:  77%|███████▋  | 7740/10000 [00:20<00:05, 410.74it/s]

Phase 1:  78%|███████▊  | 7782/10000 [00:20<00:05, 409.96it/s]

Phase 1:  78%|███████▊  | 7824/10000 [00:20<00:05, 405.78it/s]

Phase 1:  79%|███████▊  | 7865/10000 [00:20<00:05, 399.72it/s]

Phase 1:  79%|███████▉  | 7907/10000 [00:20<00:05, 403.40it/s]

Phase 1:  79%|███████▉  | 7948/10000 [00:20<00:05, 397.08it/s]

Phase 1:  80%|███████▉  | 7988/10000 [00:20<00:05, 393.17it/s]

  Step 8000: loss = 0.019666

Phase 1:  80%|████████  | 8028/10000 [00:21<00:05, 393.25it/s]

Phase 1:  81%|████████  | 8068/10000 [00:21<00:04, 392.03it/s]

Phase 1:  81%|████████  | 8108/10000 [00:21<00:04, 389.08it/s]

Phase 1:  81%|████████▏ | 8147/10000 [00:21<00:04, 381.88it/s]

Phase 1:  82%|████████▏ | 8188/10000 [00:21<00:04, 388.74it/s]

Phase 1:  82%|████████▏ | 8229/10000 [00:21<00:04, 394.50it/s]

Phase 1:  83%|████████▎ | 8271/10000 [00:21<00:04, 400.15it/s]

Phase 1:  83%|████████▎ | 8314/10000 [00:21<00:04, 406.47it/s]

Phase 1:  84%|████████▎ | 8356/10000 [00:21<00:04, 408.84it/s]

Phase 1:  84%|████████▍ | 8399/10000 [00:22<00:03, 412.78it/s]

Phase 1:  84%|████████▍ | 8441/10000 [00:22<00:03, 411.86it/s]

Phase 1:  85%|████████▍ | 8484/10000 [00:22<00:03, 414.68it/s]

Phase 1:  85%|████████▌ | 8527/10000 [00:22<00:03, 416.28it/s]

Phase 1:  86%|████████▌ | 8569/10000 [00:22<00:03, 417.09it/s]

Phase 1:  86%|████████▌ | 8611/10000 [00:22<00:03, 411.57it/s]

Phase 1:  87%|████████▋ | 8653/10000 [00:22<00:03, 408.71it/s]

Phase 1:  87%|████████▋ | 8694/10000 [00:22<00:03, 390.17it/s]

Phase 1:  87%|████████▋ | 8734/10000 [00:22<00:03, 385.38it/s]

Phase 1:  88%|████████▊ | 8773/10000 [00:23<00:03, 307.00it/s]

Phase 1:  88%|████████▊ | 8807/10000 [00:23<00:04, 266.37it/s]

Phase 1:  88%|████████▊ | 8846/10000 [00:23<00:03, 294.43it/s]

Phase 1:  89%|████████▉ | 8880/10000 [00:23<00:03, 303.74it/s]

Phase 1:  89%|████████▉ | 8913/10000 [00:23<00:04, 252.77it/s]

Phase 1:  89%|████████▉ | 8949/10000 [00:23<00:03, 276.54it/s]

Phase 1:  90%|████████▉ | 8988/10000 [00:23<00:03, 304.56it/s]

  Step 9000: loss = 0.018334

Phase 1:  90%|█████████ | 9021/10000 [00:23<00:03, 264.64it/s]

Phase 1:  91%|█████████ | 9061/10000 [00:24<00:03, 296.41it/s]

Phase 1:  91%|█████████ | 9099/10000 [00:24<00:02, 317.61it/s]

Phase 1:  91%|█████████▏| 9141/10000 [00:24<00:02, 343.16it/s]

Phase 1:  92%|█████████▏| 9184/10000 [00:24<00:02, 364.72it/s]

Phase 1:  92%|█████████▏| 9226/10000 [00:24<00:02, 379.01it/s]

Phase 1:  93%|█████████▎| 9266/10000 [00:24<00:01, 384.15it/s]

Phase 1:  93%|█████████▎| 9306/10000 [00:24<00:01, 378.51it/s]

Phase 1:  93%|█████████▎| 9345/10000 [00:24<00:01, 376.25it/s]

Phase 1:  94%|█████████▍| 9387/10000 [00:24<00:01, 388.21it/s]

Phase 1:  94%|█████████▍| 9427/10000 [00:24<00:01, 389.23it/s]

Phase 1:  95%|█████████▍| 9467/10000 [00:25<00:01, 357.41it/s]

Phase 1:  95%|█████████▌| 9504/10000 [00:25<00:01, 312.05it/s]

Phase 1:  95%|█████████▌| 9542/10000 [00:25<00:01, 328.34it/s]

Phase 1:  96%|█████████▌| 9584/10000 [00:25<00:01, 351.17it/s]

Phase 1:  96%|█████████▌| 9621/10000 [00:25<00:01, 342.29it/s]

Phase 1:  97%|█████████▋| 9657/10000 [00:25<00:01, 336.43it/s]

Phase 1:  97%|█████████▋| 9692/10000 [00:25<00:00, 331.44it/s]

Phase 1:  97%|█████████▋| 9728/10000 [00:25<00:00, 336.98it/s]

Phase 1:  98%|█████████▊| 9762/10000 [00:26<00:00, 317.22it/s]

Phase 1:  98%|█████████▊| 9805/10000 [00:26<00:00, 347.74it/s]

Phase 1:  98%|█████████▊| 9848/10000 [00:26<00:00, 368.15it/s]

Phase 1:  99%|█████████▉| 9886/10000 [00:26<00:00, 337.40it/s]

Phase 1:  99%|█████████▉| 9924/10000 [00:26<00:00, 347.12it/s]

Phase 1: 100%|█████████▉| 9960/10000 [00:26<00:00, 348.99it/s]

Phase 1: 100%|█████████▉| 9997/10000 [00:26<00:00, 353.30it/s]

Phase 1: 100%|██████████| 10000/10000 [00:26<00:00, 374.50it/s]

Phase 1 complete. Final loss: 0.018334

6. Phase 2: Supervisor Training

Train Supervisor to predict next latent step: $\mathcal{L}_S = ||h_{t+1} - \hat{s}_t||_2$

In [21]python · cell 31

python

if not SKIP_TRAINING:
    print("\n" + "=" * 60)
    print("PHASE 2: Supervisor Training")
    print("=" * 60)

    opt_supervisor = optim.Adam(supervisor.parameters(), lr=LEARNING_RATE)
    supervisor_losses = []

    for step in tqdm(range(TRAIN_STEPS), desc="Phase 2"):
        (batch,) = next(data_iter)
        batch = batch.to(device)

        with torch.no_grad():
            h = embedder(batch)

        h_hat = supervisor(h)
        g_loss_s = mse_loss(h[:, 1:, :], h_hat[:, :-1, :])

        opt_supervisor.zero_grad()
        g_loss_s.backward()
        opt_supervisor.step()

        if step % 1000 == 0:
            supervisor_losses.append(g_loss_s.item())
            print(f"  Step {step}: loss = {g_loss_s.item():.6f}", flush=True)

    print(f"Phase 2 complete. Final loss: {supervisor_losses[-1]:.6f}")

Output


============================================================
PHASE 2: Supervisor Training
============================================================

Phase 2:   0%|          | 0/10000 [00:00<?, ?it/s]

  Step 0: loss = 0.062222

Phase 2:   1%|          | 56/10000 [00:00<00:17, 553.19it/s]

Phase 2:   1%|          | 112/10000 [00:00<00:18, 545.18it/s]

Phase 2:   2%|▏         | 181/10000 [00:00<00:16, 607.73it/s]

Phase 2:   2%|▏         | 246/10000 [00:00<00:15, 623.54it/s]

Phase 2:   3%|▎         | 310/10000 [00:00<00:15, 627.31it/s]

Phase 2:   4%|▎         | 373/10000 [00:00<00:15, 607.42it/s]

Phase 2:   4%|▍         | 434/10000 [00:00<00:16, 585.27it/s]

Phase 2:   5%|▌         | 502/10000 [00:00<00:15, 613.79it/s]

Phase 2:   6%|▌         | 574/10000 [00:00<00:14, 645.35it/s]

Phase 2:   6%|▋         | 643/10000 [00:01<00:14, 657.94it/s]

Phase 2:   7%|▋         | 710/10000 [00:01<00:14, 637.08it/s]

Phase 2:   8%|▊         | 776/10000 [00:01<00:14, 640.30it/s]

Phase 2:   8%|▊         | 841/10000 [00:01<00:14, 620.16it/s]

Phase 2:   9%|▉         | 914/10000 [00:01<00:13, 649.67it/s]

Phase 2:  10%|▉         | 986/10000 [00:01<00:13, 668.12it/s]

  Step 1000: loss = 0.000380

Phase 2:  11%|█         | 1054/10000 [00:01<00:13, 660.91it/s]

Phase 2:  11%|█▏        | 1125/10000 [00:01<00:13, 673.03it/s]

Phase 2:  12%|█▏        | 1193/10000 [00:01<00:13, 672.93it/s]

Phase 2:  13%|█▎        | 1261/10000 [00:01<00:12, 674.57it/s]

Phase 2:  13%|█▎        | 1329/10000 [00:02<00:12, 672.55it/s]

Phase 2:  14%|█▍        | 1397/10000 [00:02<00:12, 671.61it/s]

Phase 2:  15%|█▍        | 1465/10000 [00:02<00:12, 667.20it/s]

Phase 2:  15%|█▌        | 1533/10000 [00:02<00:12, 669.81it/s]

Phase 2:  16%|█▌        | 1601/10000 [00:02<00:12, 670.65it/s]

Phase 2:  17%|█▋        | 1670/10000 [00:02<00:12, 674.32it/s]

Phase 2:  17%|█▋        | 1738/10000 [00:02<00:12, 651.86it/s]

Phase 2:  18%|█▊        | 1806/10000 [00:02<00:12, 658.84it/s]

Phase 2:  19%|█▊        | 1874/10000 [00:02<00:12, 664.06it/s]

Phase 2:  19%|█▉        | 1943/10000 [00:02<00:12, 669.31it/s]

  Step 2000: loss = 0.000161

Phase 2:  20%|██        | 2011/10000 [00:03<00:11, 669.96it/s]

Phase 2:  21%|██        | 2079/10000 [00:03<00:11, 671.70it/s]

Phase 2:  21%|██▏       | 2147/10000 [00:03<00:11, 673.08it/s]

Phase 2:  22%|██▏       | 2215/10000 [00:03<00:11, 673.61it/s]

Phase 2:  23%|██▎       | 2283/10000 [00:03<00:11, 674.96it/s]

Phase 2:  24%|██▎       | 2351/10000 [00:03<00:11, 674.93it/s]

Phase 2:  24%|██▍       | 2419/10000 [00:03<00:11, 675.19it/s]

Phase 2:  25%|██▍       | 2487/10000 [00:03<00:11, 646.62it/s]

Phase 2:  26%|██▌       | 2552/10000 [00:03<00:11, 623.97it/s]

Phase 2:  26%|██▌       | 2615/10000 [00:04<00:12, 613.35it/s]

Phase 2:  27%|██▋       | 2677/10000 [00:04<00:12, 605.06it/s]

Phase 2:  27%|██▋       | 2743/10000 [00:04<00:11, 619.58it/s]

Phase 2:  28%|██▊       | 2806/10000 [00:04<00:12, 560.85it/s]

Phase 2:  29%|██▊       | 2872/10000 [00:04<00:12, 585.50it/s]

Phase 2:  29%|██▉       | 2939/10000 [00:04<00:11, 608.93it/s]

  Step 3000: loss = 0.000120

Phase 2:  30%|███       | 3008/10000 [00:04<00:11, 630.46it/s]

Phase 2:  31%|███       | 3076/10000 [00:04<00:10, 644.25it/s]

Phase 2:  31%|███▏      | 3141/10000 [00:04<00:10, 642.17it/s]

Phase 2:  32%|███▏      | 3209/10000 [00:04<00:10, 653.25it/s]

Phase 2:  33%|███▎      | 3280/10000 [00:05<00:10, 669.90it/s]

Phase 2:  34%|███▎      | 3352/10000 [00:05<00:09, 683.31it/s]

Phase 2:  34%|███▍      | 3423/10000 [00:05<00:09, 691.15it/s]

Phase 2:  35%|███▍      | 3495/10000 [00:05<00:09, 697.49it/s]

Phase 2:  36%|███▌      | 3567/10000 [00:05<00:09, 702.49it/s]

Phase 2:  36%|███▋      | 3639/10000 [00:05<00:09, 705.61it/s]

Phase 2:  37%|███▋      | 3711/10000 [00:05<00:08, 708.91it/s]

Phase 2:  38%|███▊      | 3782/10000 [00:05<00:09, 683.54it/s]

Phase 2:  39%|███▊      | 3851/10000 [00:05<00:09, 682.67it/s]

Phase 2:  39%|███▉      | 3920/10000 [00:06<00:08, 681.18it/s]

Phase 2:  40%|███▉      | 3989/10000 [00:06<00:08, 679.61it/s]

  Step 4000: loss = 0.000089

Phase 2:  41%|████      | 4059/10000 [00:06<00:08, 685.21it/s]

Phase 2:  41%|████▏     | 4131/10000 [00:06<00:08, 693.62it/s]

Phase 2:  42%|████▏     | 4201/10000 [00:06<00:08, 682.18it/s]

Phase 2:  43%|████▎     | 4270/10000 [00:06<00:08, 682.13it/s]

Phase 2:  43%|████▎     | 4339/10000 [00:06<00:08, 681.12it/s]

Phase 2:  44%|████▍     | 4408/10000 [00:06<00:08, 679.26it/s]

Phase 2:  45%|████▍     | 4476/10000 [00:06<00:08, 678.96it/s]

Phase 2:  45%|████▌     | 4544/10000 [00:06<00:08, 678.56it/s]

Phase 2:  46%|████▌     | 4612/10000 [00:07<00:07, 678.85it/s]

Phase 2:  47%|████▋     | 4681/10000 [00:07<00:07, 679.22it/s]

Phase 2:  47%|████▋     | 4749/10000 [00:07<00:07, 677.63it/s]

Phase 2:  48%|████▊     | 4817/10000 [00:07<00:07, 677.85it/s]

Phase 2:  49%|████▉     | 4885/10000 [00:07<00:07, 677.97it/s]

Phase 2:  50%|████▉     | 4953/10000 [00:07<00:07, 675.97it/s]

  Step 5000: loss = 0.000087

Phase 2:  50%|█████     | 5021/10000 [00:07<00:07, 667.37it/s]

Phase 2:  51%|█████     | 5089/10000 [00:07<00:07, 668.32it/s]

Phase 2:  52%|█████▏    | 5156/10000 [00:07<00:07, 665.73it/s]

Phase 2:  52%|█████▏    | 5223/10000 [00:07<00:07, 648.43it/s]

Phase 2:  53%|█████▎    | 5291/10000 [00:08<00:07, 657.13it/s]

Phase 2:  54%|█████▎    | 5359/10000 [00:08<00:07, 661.86it/s]

Phase 2:  54%|█████▍    | 5427/10000 [00:08<00:06, 666.34it/s]

Phase 2:  55%|█████▍    | 5495/10000 [00:08<00:06, 670.06it/s]

Phase 2:  56%|█████▌    | 5563/10000 [00:08<00:06, 672.94it/s]

Phase 2:  56%|█████▋    | 5631/10000 [00:08<00:06, 662.80it/s]

Phase 2:  57%|█████▋    | 5698/10000 [00:08<00:06, 641.37it/s]

Phase 2:  58%|█████▊    | 5767/10000 [00:08<00:06, 653.62it/s]

Phase 2:  58%|█████▊    | 5833/10000 [00:08<00:06, 650.46it/s]

Phase 2:  59%|█████▉    | 5900/10000 [00:08<00:06, 654.47it/s]

Phase 2:  60%|█████▉    | 5966/10000 [00:09<00:08, 482.59it/s]

  Step 6000: loss = 0.000072

Phase 2:  60%|██████    | 6030/10000 [00:09<00:07, 519.11it/s]

Phase 2:  61%|██████    | 6088/10000 [00:09<00:07, 532.13it/s]

Phase 2:  62%|██████▏   | 6157/10000 [00:09<00:06, 571.73it/s]

Phase 2:  62%|██████▏   | 6225/10000 [00:09<00:06, 599.77it/s]

Phase 2:  63%|██████▎   | 6296/10000 [00:09<00:05, 628.96it/s]

Phase 2:  64%|██████▎   | 6361/10000 [00:09<00:06, 597.02it/s]

Phase 2:  64%|██████▍   | 6425/10000 [00:09<00:05, 605.08it/s]

Phase 2:  65%|██████▍   | 6487/10000 [00:10<00:05, 591.65it/s]

Phase 2:  66%|██████▌   | 6558/10000 [00:10<00:05, 624.26it/s]

Phase 2:  66%|██████▌   | 6623/10000 [00:10<00:05, 631.27it/s]

Phase 2:  67%|██████▋   | 6693/10000 [00:10<00:05, 649.21it/s]

Phase 2:  68%|██████▊   | 6759/10000 [00:10<00:05, 543.86it/s]

Phase 2:  68%|██████▊   | 6826/10000 [00:10<00:05, 574.55it/s]

Phase 2:  69%|██████▉   | 6895/10000 [00:10<00:05, 603.85it/s]

Phase 2:  70%|██████▉   | 6964/10000 [00:10<00:04, 625.43it/s]

  Step 7000: loss = 0.000077

Phase 2:  70%|███████   | 7029/10000 [00:10<00:04, 629.08it/s]

Phase 2:  71%|███████   | 7094/10000 [00:11<00:04, 633.28it/s]

Phase 2:  72%|███████▏  | 7165/10000 [00:11<00:04, 653.63it/s]

Phase 2:  72%|███████▏  | 7232/10000 [00:11<00:04, 658.21it/s]

Phase 2:  73%|███████▎  | 7299/10000 [00:11<00:04, 652.46it/s]

Phase 2:  74%|███████▎  | 7366/10000 [00:11<00:04, 653.64it/s]

Phase 2:  74%|███████▍  | 7432/10000 [00:11<00:04, 560.72it/s]

Phase 2:  75%|███████▍  | 7491/10000 [00:11<00:04, 504.59it/s]

Phase 2:  76%|███████▌  | 7554/10000 [00:11<00:04, 534.77it/s]

Phase 2:  76%|███████▌  | 7616/10000 [00:11<00:04, 556.67it/s]

Phase 2:  77%|███████▋  | 7674/10000 [00:12<00:04, 562.22it/s]

Phase 2:  77%|███████▋  | 7742/10000 [00:12<00:03, 594.58it/s]

Phase 2:  78%|███████▊  | 7809/10000 [00:12<00:03, 614.97it/s]

Phase 2:  79%|███████▊  | 7872/10000 [00:12<00:03, 582.67it/s]

Phase 2:  79%|███████▉  | 7938/10000 [00:12<00:03, 603.01it/s]

  Step 8000: loss = 0.000060

Phase 2:  80%|████████  | 8009/10000 [00:12<00:03, 631.87it/s]

Phase 2:  81%|████████  | 8074/10000 [00:12<00:03, 635.77it/s]

Phase 2:  81%|████████▏ | 8139/10000 [00:12<00:02, 628.04it/s]

Phase 2:  82%|████████▏ | 8208/10000 [00:12<00:02, 643.35it/s]

Phase 2:  83%|████████▎ | 8273/10000 [00:12<00:02, 635.49it/s]

Phase 2:  83%|████████▎ | 8337/10000 [00:13<00:03, 512.18it/s]

Phase 2:  84%|████████▍ | 8403/10000 [00:13<00:02, 548.43it/s]

Phase 2:  85%|████████▍ | 8472/10000 [00:13<00:02, 585.67it/s]

Phase 2:  85%|████████▌ | 8541/10000 [00:13<00:02, 612.59it/s]

Phase 2:  86%|████████▌ | 8611/10000 [00:13<00:02, 636.87it/s]

Phase 2:  87%|████████▋ | 8677/10000 [00:13<00:02, 624.31it/s]

Phase 2:  87%|████████▋ | 8741/10000 [00:13<00:02, 610.64it/s]

Phase 2:  88%|████████▊ | 8803/10000 [00:13<00:02, 554.96it/s]

Phase 2:  89%|████████▊ | 8874/10000 [00:14<00:01, 595.12it/s]

Phase 2:  89%|████████▉ | 8939/10000 [00:14<00:01, 609.22it/s]

  Step 9000: loss = 0.000059

Phase 2:  90%|█████████ | 9007/10000 [00:14<00:01, 627.79it/s]

Phase 2:  91%|█████████ | 9075/10000 [00:14<00:01, 642.75it/s]

Phase 2:  91%|█████████▏| 9143/10000 [00:14<00:01, 652.08it/s]

Phase 2:  92%|█████████▏| 9211/10000 [00:14<00:01, 659.40it/s]

Phase 2:  93%|█████████▎| 9279/10000 [00:14<00:01, 663.78it/s]

Phase 2:  93%|█████████▎| 9347/10000 [00:14<00:00, 666.86it/s]

Phase 2:  94%|█████████▍| 9414/10000 [00:14<00:00, 666.85it/s]

Phase 2:  95%|█████████▍| 9482/10000 [00:14<00:00, 670.29it/s]

Phase 2:  96%|█████████▌| 9550/10000 [00:15<00:00, 659.23it/s]

Phase 2:  96%|█████████▌| 9617/10000 [00:15<00:00, 654.48it/s]

Phase 2:  97%|█████████▋| 9684/10000 [00:15<00:00, 658.03it/s]

Phase 2:  98%|█████████▊| 9752/10000 [00:15<00:00, 663.10it/s]

Phase 2:  98%|█████████▊| 9820/10000 [00:15<00:00, 666.97it/s]

Phase 2:  99%|█████████▉| 9889/10000 [00:15<00:00, 672.68it/s]

Phase 2: 100%|█████████▉| 9958/10000 [00:15<00:00, 674.81it/s]

Phase 2: 100%|██████████| 10000/10000 [00:15<00:00, 637.59it/s]

Phase 2 complete. Final loss: 0.000059

7. Phase 3: Joint Adversarial Training

Train Generator and Discriminator adversarially while maintaining reconstruction and supervised losses. Key details:

2 generator+embedder updates per discriminator update
Discriminator gating: only train if loss > 0.15

In [22]python · cell 33

python

def get_moment_loss(y_true: torch.Tensor, y_pred: torch.Tensor) -> torch.Tensor:
    """Match first two moments between real and synthetic."""
    y_true_mean = y_true.mean(dim=0)
    y_pred_mean = y_pred.mean(dim=0)
    y_true_var = y_true.var(dim=0)
    y_pred_var = y_pred.var(dim=0)

    loss_mean = torch.abs(y_true_mean - y_pred_mean).mean()
    loss_var = torch.abs(torch.sqrt(y_true_var + 1e-6) - torch.sqrt(y_pred_var + 1e-6)).mean()

    return loss_mean + loss_var

In [23]python · cell 34

python

if not SKIP_TRAINING:
    print("\n" + "=" * 60)
    print("PHASE 3: Joint Adversarial Training")
    print("=" * 60)

    opt_generator = optim.Adam(
        list(generator.parameters()) + list(supervisor.parameters()),
        lr=LEARNING_RATE,
    )
    opt_discriminator = optim.Adam(discriminator.parameters(), lr=LEARNING_RATE)
    opt_embedder = optim.Adam(
        list(embedder.parameters()) + list(recovery.parameters()),
        lr=LEARNING_RATE,
    )

    bce_loss = nn.BCELoss()
    gamma = 1.0
    g_losses, d_losses = [], []

Output


============================================================
PHASE 3: Joint Adversarial Training
============================================================

In [24]python · cell 35

python

if not SKIP_TRAINING:
    for step in range(TRAIN_STEPS):
        # 2 generator+embedder updates per discriminator update
        for _ in range(2):
            (batch,) = next(data_iter)
            batch = batch.to(device)
            batch_size_actual = batch.shape[0]

            z = torch.rand(batch_size_actual, SEQ_LEN, n_features, device=device)

            # Generator step
            opt_generator.zero_grad()

            # Generate synthetic latent sequences
            e_hat = generator(z)
            h_hat = supervisor(e_hat)

            # Supervised loss on SYNTHETIC: generator learns temporal coherence
            # Predict e_hat[t+1] from e_hat[t], so align e_hat[1:] with h_hat[:-1]
            g_loss_s = mse_loss(e_hat[:, 1:, :], h_hat[:, :-1, :])

            y_fake = discriminator(h_hat)
            y_fake_e = discriminator(e_hat)

            g_loss_u = bce_loss(y_fake, torch.ones_like(y_fake))
            g_loss_u_e = bce_loss(y_fake_e, torch.ones_like(y_fake_e))

            x_hat = recovery(h_hat)
            g_loss_v = get_moment_loss(batch, x_hat)

            g_loss = g_loss_u + g_loss_u_e + 100.0 * torch.sqrt(g_loss_s) + 100.0 * g_loss_v
            g_loss.backward()
            opt_generator.step()

            # Embedder step
            opt_embedder.zero_grad()

            h = embedder(batch)
            h_hat_sup = supervisor(h)
            # Supervised loss: predict h[t+1] from h[t], so align h[1:] with h_hat_sup[:-1]
            g_loss_s = mse_loss(h[:, 1:, :], h_hat_sup[:, :-1, :])

            x_tilde = recovery(h)
            e_loss_t0 = mse_loss(x_tilde, batch)

            e_loss = 10.0 * torch.sqrt(e_loss_t0) + 0.1 * g_loss_s
            e_loss.backward()
            opt_embedder.step()

        # Discriminator step (with gating)
        (batch,) = next(data_iter)
        batch = batch.to(device)
        batch_size_actual = batch.shape[0]
        z = torch.rand(batch_size_actual, SEQ_LEN, n_features, device=device)

        with torch.no_grad():
            h = embedder(batch)
            e_hat = generator(z)
            h_hat = supervisor(e_hat)

        y_real = discriminator(h)
        y_fake = discriminator(h_hat)
        y_fake_e = discriminator(e_hat)

        d_loss_real = bce_loss(y_real, torch.ones_like(y_real))
        d_loss_fake = bce_loss(y_fake, torch.zeros_like(y_fake))
        d_loss_fake_e = bce_loss(y_fake_e, torch.zeros_like(y_fake_e))

        d_loss = d_loss_real + d_loss_fake + gamma * d_loss_fake_e

        # Gating: only train if loss > 0.15
        if d_loss.item() > 0.15:
            opt_discriminator.zero_grad()
            d_loss.backward()
            opt_discriminator.step()

        if step % 1000 == 0:
            g_losses.append(g_loss.item())
            d_losses.append(d_loss.item())
            print(
                f"  Step {step:6,}: D={d_loss.item():.4f}, G={g_loss.item():.4f}, "
                f"g_s={g_loss_s.item():.4f}, g_v={g_loss_v.item():.4f}",
                flush=True,
            )

    print(f"Phase 3 complete. Final D={d_losses[-1]:.4f}, G={g_losses[-1]:.4f}")

Output

  Step      0: D=2.1530, G=43.4782, g_s=0.0011, g_v=0.2834

  Step  1,000: D=1.0556, G=10.1095, g_s=0.0102, g_v=0.0450

  Step  2,000: D=1.5774, G=8.9707, g_s=0.0057, g_v=0.0528

  Step  3,000: D=1.4850, G=7.9956, g_s=0.0047, g_v=0.0378

  Step  4,000: D=1.4734, G=7.6414, g_s=0.0037, g_v=0.0360

  Step  5,000: D=1.1408, G=10.1235, g_s=0.0031, g_v=0.0469

  Step  6,000: D=1.5352, G=10.0777, g_s=0.0041, g_v=0.0476

  Step  7,000: D=1.1690, G=7.7203, g_s=0.0041, g_v=0.0305

  Step  8,000: D=1.5861, G=7.8996, g_s=0.0024, g_v=0.0363

  Step  9,000: D=1.2711, G=9.0630, g_s=0.0022, g_v=0.0435

Phase 3 complete. Final D=1.2711, G=9.0630

8. Generate Synthetic Data

In [25]python · cell 37

python

print("\n=== Generating Synthetic Data ===")

generator.eval()
supervisor.eval()
recovery.eval()

n_synthetic = len(sequences)
generated_batches = []

with torch.no_grad():
    for i in range(0, n_synthetic, BATCH_SIZE):
        batch_size = min(BATCH_SIZE, n_synthetic - i)
        z = torch.rand(batch_size, SEQ_LEN, n_features, device=device)
        e_hat = generator(z)
        h_hat = supervisor(e_hat)
        x_hat = recovery(h_hat)
        generated_batches.append(x_hat.cpu().numpy())

synthetic = np.vstack(generated_batches)
print(f"Generated {len(synthetic)} synthetic sequences")
print(f"Synthetic mean: {synthetic.mean():.4f} (real: {sequences.mean():.4f})")
print(f"Synthetic std:  {synthetic.std():.4f} (real: {sequences.std():.4f})")

Output


=== Generating Synthetic Data ===
Generated 3644 synthetic sequences
Synthetic mean: 0.3761 (real: 0.3660)
Synthetic std:  0.2532 (real: 0.2496)

9. Evaluation

We evaluate using the Fidelity-Utility-Privacy framework with LSTM-based evaluation matching the original paper.

9.1 Diversity: PCA and t-SNE Visualization

In [26]python · cell 40

python

fig = plot_fidelity_comparison(
    sequences, synthetic, title="TimeGAN: Real vs Synthetic Distribution", n_samples=1000
)
plt.show()

Output

utils/style.py:764: UserWarning: The figure layout has changed to tight
  plt.tight_layout()

<Figure size 1200x500 with 2 Axes>

[省略较大 image/png 输出]

9.2 Paper Evaluation Suite (LSTM-based)

Run the full evaluation following Yoon et al. (2019) using LSTM predictors and discriminators, not tree-based models.

In [27]python · cell 42

python

print("=" * 70)
print("TIMEGAN EVALUATION SUITE")
print("=" * 70)

# Evaluation settings per EXPERIMENTS.md:
# - hidden_dim=input_dim (small classifier to avoid overfitting)
# - epochs_disc=250 (matches benchmark)
eval_results = run_timegan_evaluation(
    synthetic=synthetic,
    real_train=sequences,
    real_holdout=holdout_sequences,
    hidden_dim=n_features,  # Must match input_dim for fair evaluation
    epochs_disc=250,  # Per benchmark experiments
    quick_test=(TRAIN_STEPS < 1000),
    verbose=True,
    include_yoon=True,
)

disc_accuracy = eval_results["discriminative"]["accuracy"]
tstr_ratio = eval_results["predictive"]["ratio"]

print("\n" + "=" * 70)
print("SUMMARY")
print("=" * 70)
print(f"Discriminative Accuracy: {disc_accuracy:.1%} (target: ~50%)")
print(f"TSTR Ratio: {tstr_ratio:.3f} (target: ~1.0)")

Output

======================================================================
TIMEGAN EVALUATION SUITE
======================================================================
============================================================
TimeGAN Evaluation Suite
============================================================

Reference: Yoon et al., NeurIPS 2019, Section 5.2

1. YOON PREDICTIVE SCORE
----------------------------------------
(Last-feature prediction per Yoon et al.)
Predictive Score (Yoon et al.)
  Features: 5 input → 1 output (last feature)
  Iterations: 5000, Batch size: 128

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.272]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.249]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.259]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.275]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.269]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.266]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.268]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.237]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.261]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.261]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.258]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.268]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.265]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.271]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.277]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.255]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.262]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.264]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.264]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.252]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.251]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.269]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.239]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.257]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.231]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.264]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.263]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.243]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.275]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.249]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.245]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.27]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.263]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.249]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.26]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.257]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.232]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.245]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.238]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.252]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.239]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.25]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.238]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.245]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.229]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.242]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.24]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.241]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.249]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.231]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.246]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.23]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.248]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.237]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.238]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.248]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.252]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.246]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.231]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.237]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.241]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.233]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.236]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.242]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.228]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.263]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.222]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.236]

Yoon Predictive:   0%|          | 0/5000 [00:00<?, ?it/s, loss=0.231]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.231]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.224]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.247]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.232]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.229]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.238]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.243]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.227]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.223]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.217]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.221]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.226]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.247]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.237]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.221]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.21]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.223]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.229]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.246]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.239]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.225]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.229]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.228]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.232]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.228]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.231]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.224]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.226]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.233]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.23]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.205]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.212]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.212]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.222]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.208]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.206]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.207]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.201]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.202]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.227]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.21]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.218]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.227]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.22]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.232]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.222]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.236]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.229]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.219]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.209]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.203]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.216]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.201]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.208]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.229]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.206]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.22]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.227]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.197]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.216]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.195]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.203]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.197]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.21]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.213]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.213]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.188]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.202]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.189]

Yoon Predictive:   1%|▏         | 69/5000 [00:00<00:07, 686.75it/s, loss=0.198]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.198]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.183]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.207]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.188]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.202]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.182]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.198]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.198]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.191]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.185]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.191]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.187]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.203]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.202]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.206]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.189]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.175]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.181]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.197]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.178]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.173]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.194]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.179]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.177]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.175]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.187]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.185]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.18]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.17]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.178]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.183]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.17]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.176]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.166]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.172]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.157]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.161]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.169]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.168]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.163]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.159]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.158]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.155]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.162]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.148]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.154]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.167]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.15]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.155]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.146]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.152]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.15]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.165]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.146]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.151]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.146]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.135]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.145]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.14]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.145]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.138]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.141]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.133]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.137]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.123]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.149]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.126]

Yoon Predictive:   3%|▎         | 138/5000 [00:00<00:07, 644.21it/s, loss=0.121]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.121]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.125]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.129]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.128]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.14]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.124]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.122]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.12]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.119]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.12]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.118]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.116]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.116]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.113]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.125]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.109]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.112]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.117]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.109]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.111]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.114]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.113]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.11]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.0984]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.103]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.098]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.11]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.114]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.0909]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.103]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.0997]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.104]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.0943]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.108]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.11]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.0999]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.0947]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.0932]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.105]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.0837]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.0947]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.0889]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.095]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.0868]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.0908]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.0885]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.0888]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.0863]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.0868]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.0844]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.09]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.0821]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.0937]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.0881]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.0847]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.0878]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.0893]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.0812]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.0861]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.0839]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.0828]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.0816]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.0887]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.0855]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.0826]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.0921]

Yoon Predictive:   4%|▍         | 205/5000 [00:00<00:07, 654.55it/s, loss=0.0782]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0782]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0868]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0821]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0804]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0791]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0775]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0804]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0802]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0807]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0793]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0817]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0742]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0776]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0799]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0792]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0762]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0733]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0816]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0778]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0777]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0758]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0708]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0779]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.074]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0759]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0692]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0779]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0759]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0717]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0735]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0668]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0709]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0707]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0724]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0669]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0671]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0729]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0669]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0682]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0767]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0657]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0719]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0678]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0637]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0685]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0711]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0677]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0673]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0647]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0685]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0722]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0651]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0695]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.069]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0728]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0662]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0636]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0677]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0635]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.068]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0681]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0632]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0624]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0618]

Yoon Predictive:   5%|▌         | 271/5000 [00:00<00:07, 632.98it/s, loss=0.0691]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0691]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0694]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0675]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0606]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0631]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0644]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0677]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0662]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0678]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0637]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0652]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0627]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0638]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0602]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0654]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.064]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0641]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0638]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0603]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0672]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0596]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0642]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0639]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0647]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0603]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0617]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0585]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0615]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0623]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0593]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0683]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0633]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0658]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0615]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.057]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0625]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.056]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0556]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0592]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.058]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.059]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0672]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0574]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0603]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0569]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0594]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0547]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0553]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0606]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0605]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0619]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0585]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0538]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0584]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0552]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0572]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0528]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0566]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0564]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0559]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0582]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0538]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0557]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0508]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0535]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0563]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0581]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0616]

Yoon Predictive:   7%|▋         | 335/5000 [00:00<00:07, 631.61it/s, loss=0.0541]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0541]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0513]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0552]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0568]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0595]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0531]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0544]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0532]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0523]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0526]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0557]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.049]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0542]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0525]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.053]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0567]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0526]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0524]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0542]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0504]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0532]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0539]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0511]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0533]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0483]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0562]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.054]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0533]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0517]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0514]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0495]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0502]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0535]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0505]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.047]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0494]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0523]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0492]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0539]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0503]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0511]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0479]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0481]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0476]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0491]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0508]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.052]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0477]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0513]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0493]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0491]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0455]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0471]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0499]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0483]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0485]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0528]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0472]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0509]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0498]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0458]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0497]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0492]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0467]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0451]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0462]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0506]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0447]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0458]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0493]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0468]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0476]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0442]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0469]

Yoon Predictive:   8%|▊         | 403/5000 [00:00<00:07, 645.96it/s, loss=0.0466]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0466]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0454]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0455]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0433]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0475]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0447]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0462]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0477]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0442]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0458]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0448]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0472]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0413]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0455]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0463]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0437]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0459]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0449]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0457]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0458]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0444]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0455]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0433]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0452]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0409]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0412]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0445]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.043]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0434]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0419]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0464]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0418]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0436]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0447]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0443]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0438]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0433]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.042]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0436]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0467]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0427]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0435]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.045]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0449]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0426]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0386]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0422]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0426]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0444]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0439]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0455]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0425]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0402]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0401]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0456]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0444]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0417]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0393]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0415]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0415]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0413]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0423]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0427]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0397]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0412]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0411]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.039]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0415]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0465]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0411]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.041]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0426]

Yoon Predictive:  10%|▉         | 477/5000 [00:00<00:06, 674.83it/s, loss=0.0442]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0442]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0394]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0424]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0388]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.041]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0382]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0394]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0389]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0417]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0388]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0383]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.039]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0368]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0394]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.038]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0396]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0393]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0372]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0391]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0377]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0375]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0452]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0398]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0402]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0393]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0354]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0377]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0414]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0399]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0379]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0367]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0399]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0359]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.041]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0374]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.036]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0367]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.037]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0359]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0387]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0375]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0374]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0369]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0385]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.039]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0372]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.036]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0403]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0359]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0363]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0354]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0339]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.036]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0364]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0357]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0348]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0387]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0356]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0368]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0375]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0372]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0376]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0362]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0341]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0365]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0364]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0382]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0373]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0387]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0369]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.0379]

Yoon Predictive:  11%|█         | 549/5000 [00:00<00:06, 688.56it/s, loss=0.034]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.034]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0365]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.034]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0377]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0366]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.035]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0319]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0339]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0364]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0355]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0335]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0339]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0333]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0346]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0327]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0368]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0322]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0349]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0363]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.035]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.036]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0326]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0347]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0343]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0292]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0338]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0381]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0362]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0341]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0365]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0323]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0365]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0342]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0349]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0332]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0377]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0305]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0318]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0334]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0359]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0343]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.032]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0307]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0315]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0357]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0328]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0352]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0325]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0343]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0348]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.035]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0319]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0314]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0312]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0347]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0306]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.031]

Yoon Predictive:  12%|█▏        | 620/5000 [00:00<00:06, 694.60it/s, loss=0.0324]

Yoon Predictive:  12%|█▏        | 620/5000 [00:01<00:06, 694.60it/s, loss=0.0305]

Yoon Predictive:  12%|█▏        | 620/5000 [00:01<00:06, 694.60it/s, loss=0.0348]

Yoon Predictive:  12%|█▏        | 620/5000 [00:01<00:06, 694.60it/s, loss=0.0337]

Yoon Predictive:  12%|█▏        | 620/5000 [00:01<00:06, 694.60it/s, loss=0.033]

Yoon Predictive:  12%|█▏        | 620/5000 [00:01<00:06, 694.60it/s, loss=0.0286]

Yoon Predictive:  12%|█▏        | 620/5000 [00:01<00:06, 694.60it/s, loss=0.0325]

Yoon Predictive:  12%|█▏        | 620/5000 [00:01<00:06, 694.60it/s, loss=0.0339]

Yoon Predictive:  12%|█▏        | 620/5000 [00:01<00:06, 694.60it/s, loss=0.0266]

Yoon Predictive:  12%|█▏        | 620/5000 [00:01<00:06, 694.60it/s, loss=0.0299]

Yoon Predictive:  12%|█▏        | 620/5000 [00:01<00:06, 694.60it/s, loss=0.0317]

Yoon Predictive:  12%|█▏        | 620/5000 [00:01<00:06, 694.60it/s, loss=0.0308]

Yoon Predictive:  12%|█▏        | 620/5000 [00:01<00:06, 694.60it/s, loss=0.0332]

Yoon Predictive:  12%|█▏        | 620/5000 [00:01<00:06, 694.60it/s, loss=0.0321]

Yoon Predictive:  12%|█▏        | 620/5000 [00:01<00:06, 694.60it/s, loss=0.0347]

Yoon Predictive:  12%|█▏        | 620/5000 [00:01<00:06, 694.60it/s, loss=0.0291]

Yoon Predictive:  12%|█▏        | 620/5000 [00:01<00:06, 694.60it/s, loss=0.0316]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0316]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0313]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0334]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0302]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0323]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0299]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0303]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0313]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0313]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0285]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.034]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0345]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0326]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0288]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0296]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0327]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0303]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0305]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0337]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0318]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0299]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0329]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0303]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0266]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0288]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0308]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0304]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0292]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0304]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0284]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0291]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0298]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0318]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0294]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0298]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0269]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0272]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0294]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0284]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0281]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0286]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0267]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0262]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0325]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0302]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0315]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0267]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0316]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0299]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0302]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0288]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0271]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0288]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0274]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0291]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0277]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0305]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0284]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0257]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0265]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0267]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0276]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0288]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0253]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0292]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0281]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0272]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0301]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0283]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0279]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0269]

Yoon Predictive:  14%|█▍        | 693/5000 [00:01<00:06, 702.95it/s, loss=0.0295]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0295]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0286]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0282]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0276]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0314]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0306]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0267]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0255]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0298]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0263]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0258]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0255]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0326]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0263]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0251]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0271]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0287]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0307]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.027]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0276]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0275]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0274]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0265]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0317]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0278]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0269]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0247]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0309]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0277]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0261]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0289]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0276]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0279]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0306]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0318]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0286]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0281]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0271]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0248]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0253]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0282]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0256]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0241]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0267]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.026]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.026]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0255]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.026]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0295]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0282]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0252]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0265]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0236]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0287]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0264]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0284]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0236]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0254]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.029]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0285]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0246]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0254]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0255]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0274]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0242]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0281]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.023]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0264]

Yoon Predictive:  15%|█▌        | 764/5000 [00:01<00:06, 671.60it/s, loss=0.0272]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0272]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.025]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0253]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0247]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0263]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0298]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0257]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0254]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0252]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0262]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0272]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0265]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0247]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0258]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.028]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0285]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0254]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0279]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0274]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.022]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0277]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0272]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0238]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0235]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0242]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0268]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0251]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0266]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0256]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0235]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0261]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0237]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0217]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0247]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0237]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0267]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0268]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0263]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0258]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0306]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0297]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0297]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0268]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0241]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0282]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.026]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0238]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.026]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0247]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0251]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0263]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0248]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0265]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0251]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0246]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0294]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0238]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0237]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0237]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0263]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0239]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0262]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0259]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0253]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0239]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0261]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0273]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.024]

Yoon Predictive:  17%|█▋        | 832/5000 [00:01<00:06, 671.57it/s, loss=0.0251]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0251]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0255]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0209]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0223]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0273]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0256]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0257]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0261]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0269]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0219]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0284]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0255]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0239]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0228]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0225]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0275]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0289]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0251]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0247]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0266]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0245]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.026]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0247]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0223]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0243]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0231]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0278]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0246]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0285]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0254]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0227]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0288]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0234]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0284]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0246]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0225]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0253]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0265]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.025]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0225]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.023]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0234]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0255]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0243]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0256]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0221]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0259]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0226]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0248]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0235]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0264]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0267]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0252]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.023]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.025]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0276]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0237]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0238]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0239]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0229]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0281]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0247]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0254]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0259]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0248]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0219]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0229]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0241]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0244]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0253]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0285]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0235]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0206]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0241]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0255]

Yoon Predictive:  18%|█▊        | 900/5000 [00:01<00:06, 664.03it/s, loss=0.0256]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0256]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0228]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0276]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0261]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0224]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.023]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0239]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0266]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0215]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0228]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0229]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0276]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0264]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0233]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0235]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0245]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.027]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0272]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0277]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0256]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.026]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0243]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0212]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0237]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0255]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0235]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0232]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0225]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0282]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.024]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0268]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0237]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0234]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0244]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0243]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0269]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.025]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0224]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0241]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0245]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0275]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.024]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0249]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.024]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0223]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0265]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0212]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0237]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0244]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0242]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0269]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0252]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0221]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.022]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0267]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0224]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0228]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0207]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0269]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0223]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.027]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0253]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0219]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0222]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0219]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0251]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0266]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0251]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0255]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0222]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0225]

Yoon Predictive:  20%|█▉        | 975/5000 [00:01<00:05, 685.54it/s, loss=0.0255]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0255]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0298]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.023]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0236]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0226]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0259]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0222]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0229]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0291]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0278]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.025]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0229]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.025]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.022]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.026]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0241]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0222]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0217]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0257]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0209]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0222]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0252]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0254]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0264]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0232]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.022]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0281]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0245]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0246]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0241]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.028]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0265]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0212]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0249]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0217]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0261]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0255]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0255]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0238]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0264]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0227]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0237]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0273]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0223]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0242]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0268]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0227]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0252]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0257]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0242]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0274]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0231]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0235]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0234]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0211]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.023]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0242]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0231]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0255]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0227]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0239]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0277]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0211]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0245]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0235]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0249]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0236]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0217]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0206]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0238]

Yoon Predictive:  21%|██        | 1046/5000 [00:01<00:05, 691.52it/s, loss=0.0232]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0232]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0196]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0255]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0232]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0226]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0229]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0254]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0233]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0212]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0229]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0221]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.028]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0207]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0264]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0226]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0222]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0246]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0226]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0251]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0236]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0267]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0232]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.022]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0197]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0234]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0202]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0235]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0234]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0231]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0233]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0214]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0284]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0245]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0228]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.02]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0238]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0232]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0249]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0227]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0219]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0228]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0219]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0254]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0233]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0246]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0251]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0224]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.025]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0232]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0239]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0244]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0247]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0251]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0266]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0227]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0235]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0206]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.025]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0243]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0254]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0239]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0262]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0249]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0225]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0267]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0262]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0222]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0199]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0224]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0204]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.02]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0246]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0236]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0221]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0255]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0203]

Yoon Predictive:  22%|██▏       | 1116/5000 [00:01<00:05, 686.82it/s, loss=0.0238]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0238]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0265]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.026]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0249]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0231]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0274]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0229]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0217]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0254]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0192]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0239]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0238]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.021]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0219]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0248]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0228]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0262]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0243]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0271]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0292]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0235]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0243]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0247]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0256]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0238]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0224]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0227]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.022]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0213]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0236]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0227]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0223]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0247]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0215]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0224]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0224]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0241]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0232]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0241]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0215]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0219]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0265]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0264]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0254]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0214]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0236]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0233]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0239]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0253]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0236]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0252]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.026]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0243]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0208]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0222]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0241]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.018]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0244]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0255]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0225]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0242]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0224]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0255]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0242]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0246]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0196]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0205]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0208]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0229]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0213]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0208]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0235]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0204]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0229]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0223]

Yoon Predictive:  24%|██▍       | 1192/5000 [00:01<00:05, 706.10it/s, loss=0.0206]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0206]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.023]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0242]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0228]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0218]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0249]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0251]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0226]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.024]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0213]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0235]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.024]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0251]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.023]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0236]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0237]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0218]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0228]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0185]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0218]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0226]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0272]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0218]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0224]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.021]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0219]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0229]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0201]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0234]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0284]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0232]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0226]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.025]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0253]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0258]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0196]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0254]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0204]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0209]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0239]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0223]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0269]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0247]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.024]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0243]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0206]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0228]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0238]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0247]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0216]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0227]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0252]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0225]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0224]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0251]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0212]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0218]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0272]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0231]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0251]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0238]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.02]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0212]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0256]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0234]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0206]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0241]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.021]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0259]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0222]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0238]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0252]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0198]

Yoon Predictive:  25%|██▌       | 1267/5000 [00:01<00:05, 716.83it/s, loss=0.0234]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:01<00:05, 719.95it/s, loss=0.0234]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:01<00:05, 719.95it/s, loss=0.0246]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:01<00:05, 719.95it/s, loss=0.0191]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:01<00:05, 719.95it/s, loss=0.0243]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:01<00:05, 719.95it/s, loss=0.0249]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:01<00:05, 719.95it/s, loss=0.0238]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:01<00:05, 719.95it/s, loss=0.0241]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:01<00:05, 719.95it/s, loss=0.0226]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:01<00:05, 719.95it/s, loss=0.0222]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:01<00:05, 719.95it/s, loss=0.022]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:01<00:05, 719.95it/s, loss=0.0241]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:01<00:05, 719.95it/s, loss=0.021]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:01<00:05, 719.95it/s, loss=0.0237]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:01<00:05, 719.95it/s, loss=0.0202]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:01<00:05, 719.95it/s, loss=0.0235]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:01<00:05, 719.95it/s, loss=0.0261]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:01<00:05, 719.95it/s, loss=0.0198]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:01<00:05, 719.95it/s, loss=0.0208]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:01<00:05, 719.95it/s, loss=0.0249]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:01<00:05, 719.95it/s, loss=0.0236]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:01<00:05, 719.95it/s, loss=0.0237]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:01<00:05, 719.95it/s, loss=0.0235]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:01<00:05, 719.95it/s, loss=0.0239]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:01<00:05, 719.95it/s, loss=0.0233]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:01<00:05, 719.95it/s, loss=0.0216]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:01<00:05, 719.95it/s, loss=0.0223]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:01<00:05, 719.95it/s, loss=0.0247]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:01<00:05, 719.95it/s, loss=0.0232]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:01<00:05, 719.95it/s, loss=0.0195]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:01<00:05, 719.95it/s, loss=0.0216]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:01<00:05, 719.95it/s, loss=0.0218]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0211]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0219]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.023]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0213]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0236]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0202]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0213]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0248]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0197]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0242]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0268]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0236]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0249]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0234]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0268]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0203]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0204]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0199]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0214]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0209]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0221]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0215]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.024]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0238]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0229]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0245]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0208]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0224]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0238]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0242]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0239]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0209]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0223]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0206]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0263]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0217]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0237]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.023]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0221]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0229]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0251]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0245]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.0219]

Yoon Predictive:  27%|██▋       | 1340/5000 [00:02<00:05, 719.95it/s, loss=0.023]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.023]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0242]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0232]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0221]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0244]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0238]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0273]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0212]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0206]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0228]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0221]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0217]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.022]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0275]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0215]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.022]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0226]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0215]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0206]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.028]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0214]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0202]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0257]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0239]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0227]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0221]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0252]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0237]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0219]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0208]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.025]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0186]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0239]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0206]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0237]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.02]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0229]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.021]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0249]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0184]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.021]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0251]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0242]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0186]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0229]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0234]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0232]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0272]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0223]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0214]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0219]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0232]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.018]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0246]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0259]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0224]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.025]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.026]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0265]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0223]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0197]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0201]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0244]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0212]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0223]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0244]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0179]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0211]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0224]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0228]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0227]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0222]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0246]

Yoon Predictive:  28%|██▊       | 1414/5000 [00:02<00:04, 724.10it/s, loss=0.0233]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0233]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0192]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0272]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0227]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0271]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0229]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0229]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0232]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0272]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0235]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0208]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0204]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0236]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.019]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0182]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0209]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0226]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0224]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0216]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0222]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0194]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0214]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0222]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0182]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0232]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0199]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0261]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0212]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0218]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0247]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0232]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0222]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0209]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0191]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0221]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0215]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0228]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0196]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0206]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0192]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0268]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0208]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0228]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0204]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0217]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0219]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0204]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0192]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0196]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0241]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0199]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0248]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0208]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0263]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.02]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0245]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.022]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.022]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.02]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0219]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0235]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0223]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0222]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.022]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.02]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0233]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0239]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0221]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0206]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0252]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0266]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0203]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0248]

Yoon Predictive:  30%|██▉       | 1487/5000 [00:02<00:04, 719.88it/s, loss=0.0253]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0253]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0207]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.024]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0202]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.02]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0236]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0205]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0224]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0201]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0197]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0159]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0227]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0236]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0224]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0208]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0231]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.02]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0196]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.022]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0229]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.02]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.021]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0214]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0228]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0213]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0243]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0184]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0235]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0224]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0183]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0221]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0197]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0199]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0216]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0248]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0192]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0211]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0242]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0237]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.023]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0199]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0212]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0259]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0246]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0237]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0219]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0259]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0206]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0187]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0199]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0203]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0232]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0205]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0202]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.023]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0213]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0236]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0226]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.02]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.017]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0221]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0197]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0183]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0206]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0222]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0211]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.021]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.022]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.02]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.0211]

Yoon Predictive:  31%|███       | 1560/5000 [00:02<00:04, 688.71it/s, loss=0.02]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.02]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0206]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0186]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0179]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0204]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.022]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0211]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0236]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0237]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0257]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0221]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0231]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0196]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0213]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0242]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0182]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0219]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0231]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0208]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0231]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0243]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0214]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.018]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0204]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0216]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0209]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0197]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0226]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0197]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0243]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0248]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0215]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0223]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0198]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0227]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0212]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0209]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0203]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0226]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0265]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0215]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0215]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0217]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0212]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0195]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0243]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0221]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0205]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0183]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0209]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0193]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0237]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0218]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.02]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0218]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.021]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0234]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0202]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0217]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.023]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0243]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0216]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0212]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0196]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0213]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0252]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0185]

Yoon Predictive:  33%|███▎      | 1630/5000 [00:02<00:05, 662.83it/s, loss=0.0243]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0243]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0178]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0198]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0224]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0195]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0237]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0191]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0208]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.02]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0205]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0206]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0223]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0216]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0205]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0215]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0207]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0236]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.024]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0194]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.02]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0222]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0187]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0232]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0208]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0208]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0185]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0216]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0196]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0178]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0198]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0202]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0246]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0188]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0203]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.02]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0213]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0239]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0171]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0238]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0249]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0241]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.017]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0243]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0187]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0234]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0218]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0212]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0234]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0225]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0209]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0215]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0223]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0186]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0186]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0201]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0267]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0196]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0214]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.024]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0219]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0189]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0185]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0203]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0206]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.0181]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.023]

Yoon Predictive:  34%|███▍      | 1697/5000 [00:02<00:05, 654.67it/s, loss=0.02]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.02]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0222]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.02]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0191]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.025]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0228]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0223]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0247]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0189]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0262]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0218]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0197]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0199]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0211]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0188]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0201]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.024]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0257]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0209]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0228]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.022]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.018]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0203]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0244]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.021]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0179]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0214]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0204]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0197]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.022]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0196]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0241]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0225]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0216]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0191]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0223]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0225]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0221]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0199]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0203]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0233]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.019]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0254]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0185]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0227]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0222]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0181]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0206]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0236]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.02]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0202]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0213]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0183]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0188]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0273]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0196]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0159]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.021]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0243]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0208]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0205]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0172]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0262]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0194]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.0188]

Yoon Predictive:  35%|███▌      | 1763/5000 [00:02<00:04, 648.00it/s, loss=0.021]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.021]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0188]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0203]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0241]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0216]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0161]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0185]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0207]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0217]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.017]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0193]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0221]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0185]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0183]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0211]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0217]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0244]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0181]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.02]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.024]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0186]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0183]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0249]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0235]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0205]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0176]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0206]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0201]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0185]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0224]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0184]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0214]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0192]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0194]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0224]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0215]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0241]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0238]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0222]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0213]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0215]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0231]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0202]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0205]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0242]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0212]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0213]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.021]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0211]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0177]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0218]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0195]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0185]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0204]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0201]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0235]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0205]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0182]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0245]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.022]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0206]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0193]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0224]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0216]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0229]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0223]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0182]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0183]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0231]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0212]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.022]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0211]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0247]

Yoon Predictive:  37%|███▋      | 1828/5000 [00:02<00:04, 641.29it/s, loss=0.0219]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0219]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0161]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0191]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.021]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0198]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0178]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0201]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0248]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0215]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0182]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0214]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0196]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0158]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0169]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0192]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.024]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.019]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0202]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0268]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0175]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0245]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0173]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.021]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0186]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0236]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0195]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.023]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0215]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0218]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0229]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0249]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0195]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0215]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0195]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0171]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0227]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0211]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0203]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.019]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0184]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0182]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0188]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0207]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0235]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0201]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.019]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0201]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0167]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0208]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0203]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0191]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0226]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0191]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0183]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0201]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.022]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0253]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0204]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0178]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0218]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0209]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.018]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0188]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0165]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0195]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0225]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0211]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.021]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0209]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0256]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0204]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0195]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0206]

Yoon Predictive:  38%|███▊      | 1901/5000 [00:02<00:04, 665.10it/s, loss=0.0199]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0199]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0191]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0185]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0228]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0204]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0169]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0208]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0196]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0184]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0213]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0202]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0182]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0194]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0197]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0207]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0215]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0223]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0236]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0214]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0205]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0231]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0239]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0224]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0159]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0234]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0205]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0221]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0227]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0191]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0232]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0199]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0211]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0176]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0204]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0226]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0194]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0203]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0214]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0199]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0175]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0201]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0203]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0221]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0189]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0227]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0226]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0204]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0213]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0222]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0198]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0203]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0171]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0178]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0207]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0246]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0216]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0202]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0152]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0188]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.0161]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:02<00:04, 681.97it/s, loss=0.018]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:03<00:04, 681.97it/s, loss=0.0197]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:03<00:04, 681.97it/s, loss=0.0179]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:03<00:04, 681.97it/s, loss=0.0209]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:03<00:04, 681.97it/s, loss=0.0215]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:03<00:04, 681.97it/s, loss=0.0185]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:03<00:04, 681.97it/s, loss=0.0226]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:03<00:04, 681.97it/s, loss=0.022]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:03<00:04, 681.97it/s, loss=0.0218]

Yoon Predictive:  39%|███▉      | 1974/5000 [00:03<00:04, 681.97it/s, loss=0.0208]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0208]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0234]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0192]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0225]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0208]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0191]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0217]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0184]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.016]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0207]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0174]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0189]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0171]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0224]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0212]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0201]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0234]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0223]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0221]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0229]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0213]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0223]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0184]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0227]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0159]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0199]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0147]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0205]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.022]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0204]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.016]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0184]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0183]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0188]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0205]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0202]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0207]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0187]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0216]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0172]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0216]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0169]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0196]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.019]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0156]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0197]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0224]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0198]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0196]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0194]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0201]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0212]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.021]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0195]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0241]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0204]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0216]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0221]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0155]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0232]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0213]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0187]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0207]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0175]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0216]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0154]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0205]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0213]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0202]

Yoon Predictive:  41%|████      | 2043/5000 [00:03<00:04, 680.12it/s, loss=0.0213]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0213]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0225]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0166]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0188]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0185]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0221]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0197]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0171]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0194]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0195]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0193]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0181]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0223]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0184]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0164]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0193]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0244]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0183]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0226]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0198]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0207]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0176]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0234]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.02]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0196]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0191]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0201]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.019]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0176]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.018]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0163]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0212]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0173]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0185]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0162]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0197]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0238]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0162]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0193]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0192]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.018]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0185]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0174]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0175]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0168]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0206]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0194]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0209]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0194]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0172]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0197]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0219]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0205]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0206]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0213]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0201]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0175]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0223]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0179]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0175]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0184]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0189]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0227]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0207]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0182]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0187]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0189]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.017]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0194]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0179]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.015]

Yoon Predictive:  42%|████▏     | 2112/5000 [00:03<00:04, 679.91it/s, loss=0.0222]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0222]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0207]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0213]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0239]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0157]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0171]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0213]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0179]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0228]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0209]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0191]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0188]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0171]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.019]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0237]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0223]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0163]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0203]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0171]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0205]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0206]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0204]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.019]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0221]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0188]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0173]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0183]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0222]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0202]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.018]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0197]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0212]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0183]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0212]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0163]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0154]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0184]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0195]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0231]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0204]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0179]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0229]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.016]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0195]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0198]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0219]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0196]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0177]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.017]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0192]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0201]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0168]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0213]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0215]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0209]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0193]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0219]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0183]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.022]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0188]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0166]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0189]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0188]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.018]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0196]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0169]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0206]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0202]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0169]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.02]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0226]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0169]

Yoon Predictive:  44%|████▎     | 2183/5000 [00:03<00:04, 687.27it/s, loss=0.0188]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0188]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0191]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0219]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0202]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0179]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0202]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0184]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0229]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0136]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0228]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0188]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0206]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0177]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.02]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0184]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0178]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0168]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.021]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0197]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0181]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0193]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0161]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0166]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0179]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0199]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0194]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0196]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0204]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0185]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0158]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0187]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0166]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0157]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0181]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.019]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0206]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0219]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.019]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0178]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0177]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0166]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0171]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0193]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0198]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0195]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0209]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0198]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0185]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0182]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0212]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0174]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.018]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0189]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0199]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.015]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0198]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0202]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.022]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0177]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0221]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0181]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0182]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0183]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0173]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0186]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0206]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0166]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0176]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0228]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0198]

Yoon Predictive:  45%|████▌     | 2255/5000 [00:03<00:03, 693.28it/s, loss=0.0186]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0186]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0135]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0181]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0188]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0197]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0191]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0172]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0211]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0191]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0151]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0163]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0177]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0177]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0178]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0224]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0174]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0172]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0183]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0199]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0181]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0217]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0179]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.018]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0162]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0191]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0199]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0167]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0211]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0181]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0198]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0183]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0182]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0186]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0185]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0223]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0169]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0162]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.017]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0192]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0205]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0179]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0174]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0196]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0187]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0195]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0229]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0172]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0205]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0216]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0164]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.016]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0203]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.019]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0193]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0176]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0209]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0178]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0186]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0196]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0168]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0193]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0176]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0169]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0211]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.018]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0192]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0193]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0188]

Yoon Predictive:  46%|████▋     | 2325/5000 [00:03<00:03, 672.10it/s, loss=0.0185]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0185]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.02]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0183]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0192]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0172]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0204]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0173]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0178]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0188]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0205]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0155]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0197]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0197]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0155]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0155]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0169]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0166]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0191]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0165]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0191]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0191]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0169]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0203]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0189]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0159]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.02]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0209]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0177]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0191]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.015]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0187]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0174]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0212]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0169]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0199]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0165]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0185]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0186]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.019]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0182]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0175]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0169]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0171]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0178]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0165]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0145]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0166]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0187]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0205]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0234]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0211]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0174]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0162]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0171]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0183]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0214]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0197]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0185]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0174]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0155]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0183]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0183]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0188]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0175]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0175]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.023]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0172]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0135]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0169]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0173]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0165]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0195]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0188]

Yoon Predictive:  48%|████▊     | 2393/5000 [00:03<00:03, 670.13it/s, loss=0.0202]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0202]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0226]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0189]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0171]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0198]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0174]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0185]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.017]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0186]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0165]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0171]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0183]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0174]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0139]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0177]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0181]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0218]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0193]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0175]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0212]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0181]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0178]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0226]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0131]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0175]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.016]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0164]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.019]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0149]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0202]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0162]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0158]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0179]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0165]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0207]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0176]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0215]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.02]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0161]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0196]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0166]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0158]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0184]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0206]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0205]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0168]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0154]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0213]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.016]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.016]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0193]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0199]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0165]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0198]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.019]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0204]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0191]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0187]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0162]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0146]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0187]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0173]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0164]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0193]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.019]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0183]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0169]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0174]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0218]

Yoon Predictive:  49%|████▉     | 2466/5000 [00:03<00:03, 687.09it/s, loss=0.0184]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0184]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0167]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0177]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0148]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0203]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0194]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0164]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0186]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.018]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0158]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0181]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0171]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0186]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0169]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0173]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0133]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0195]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0137]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0203]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0169]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0192]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0175]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0192]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.016]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0183]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0184]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0194]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0162]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0202]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0176]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0156]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.016]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0202]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0197]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0167]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0185]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0183]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0197]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0176]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0178]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0171]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.017]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0169]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.017]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0166]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0208]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0187]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0159]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0152]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0157]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0189]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0159]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0164]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0181]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0191]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0151]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0164]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0185]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0192]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0162]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0175]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0169]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0227]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0173]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0176]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0184]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0177]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0175]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.0175]

Yoon Predictive:  51%|█████     | 2535/5000 [00:03<00:03, 686.96it/s, loss=0.016]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.016]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0165]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0157]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0186]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0179]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0213]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0192]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0179]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0162]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0153]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0188]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0207]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0158]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.016]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0157]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0134]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0196]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0153]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.017]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0142]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0182]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0168]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0175]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0143]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0182]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0219]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.019]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0153]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0166]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0199]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0156]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0145]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.015]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0173]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0179]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0232]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0183]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0163]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0155]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0147]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0153]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0184]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0151]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0149]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0185]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.018]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0223]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0166]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0173]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0208]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0155]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0172]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.019]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0184]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0193]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0203]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0186]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0229]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0206]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0207]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0181]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0153]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0176]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0229]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0162]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0213]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0156]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0175]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0191]

Yoon Predictive:  52%|█████▏    | 2604/5000 [00:03<00:04, 589.97it/s, loss=0.0171]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:03<00:03, 615.53it/s, loss=0.0171]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:03<00:03, 615.53it/s, loss=0.0163]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:03<00:03, 615.53it/s, loss=0.0172]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:03<00:03, 615.53it/s, loss=0.0186]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:03<00:03, 615.53it/s, loss=0.0189]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:03<00:03, 615.53it/s, loss=0.0196]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:03<00:03, 615.53it/s, loss=0.0143]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:03<00:03, 615.53it/s, loss=0.0181]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:03<00:03, 615.53it/s, loss=0.0188]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0193]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0191]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0187]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.015]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0182]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0192]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0163]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0191]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0169]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0142]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0192]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.02]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0169]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0166]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0202]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0155]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0164]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0154]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0191]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0191]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0194]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0189]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0209]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0167]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0198]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0203]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0144]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0183]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0174]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0182]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0152]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0191]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0167]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0201]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0178]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0186]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0198]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0158]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0162]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0151]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0195]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0195]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0194]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0187]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0167]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0144]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0199]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0182]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0222]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0174]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0183]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0152]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0161]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0179]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0211]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0157]

Yoon Predictive:  53%|█████▎    | 2673/5000 [00:04<00:03, 615.53it/s, loss=0.0153]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0153]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0161]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0176]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0188]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0158]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.019]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0162]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0191]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0146]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0162]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0154]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0162]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0146]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0181]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0177]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0144]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.017]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0137]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0161]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0156]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0144]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.017]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0151]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0192]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0163]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.019]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0194]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0149]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.019]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0145]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0208]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0191]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0162]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0169]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0167]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0186]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0175]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0176]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0189]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0193]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0173]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0168]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0163]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0143]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0186]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0202]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0157]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0196]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0181]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0176]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.016]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0167]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0184]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0165]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0152]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0165]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0157]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0158]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0182]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.015]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0161]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0165]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0144]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0152]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0172]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0165]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0174]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0166]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0176]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.016]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0156]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0175]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0151]

Yoon Predictive:  55%|█████▍    | 2738/5000 [00:04<00:03, 624.82it/s, loss=0.0155]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0155]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0151]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0194]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0198]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0166]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0189]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0169]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0188]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0176]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0156]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.017]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0191]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0163]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0193]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0198]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0191]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0182]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0173]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0183]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0159]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0173]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0163]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0163]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0153]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0159]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0154]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0182]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0194]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0162]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0162]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0187]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0183]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0143]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0184]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0154]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0182]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0152]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0161]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0182]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0181]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0173]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0184]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0166]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0167]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0168]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0177]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.016]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0169]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0177]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0144]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0194]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0165]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0178]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0167]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0204]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0175]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0165]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.015]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0136]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0164]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0166]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0163]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0174]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0146]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0159]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0154]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0142]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0204]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0171]

Yoon Predictive:  56%|█████▌    | 2811/5000 [00:04<00:03, 653.50it/s, loss=0.0191]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0191]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.016]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0189]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0177]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0187]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0163]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0198]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0183]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0134]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0186]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0153]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0182]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0175]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0148]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0173]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0175]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0154]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.017]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0147]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0173]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0166]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0185]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0141]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0192]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0189]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.018]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0205]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0163]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0158]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.018]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0169]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0174]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0146]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0143]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0178]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0178]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0171]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0146]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0145]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0138]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0201]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0183]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0156]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0169]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0163]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0172]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0181]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0198]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0157]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0152]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0207]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0161]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0161]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0152]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.017]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0175]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0158]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0183]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0163]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0214]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0177]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0166]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0169]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0183]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0172]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0161]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0198]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0182]

Yoon Predictive:  58%|█████▊    | 2880/5000 [00:04<00:03, 662.93it/s, loss=0.0206]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0206]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0172]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0174]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0158]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0159]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0175]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0147]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0192]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0166]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0147]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0161]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0166]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0167]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0164]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0143]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0146]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0193]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0163]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0156]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0168]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0159]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0142]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0175]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0149]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0155]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0156]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0162]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0169]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0175]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0182]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0158]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0176]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0159]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0207]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0161]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0183]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0166]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0137]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0157]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0148]

Yoon Predictive:  59%|█████▉    | 2948/5000 [00:04<00:03, 656.71it/s, loss=0.0144]

Yoon Predictive
…

[输出过长，已省略]

9.3 Training Curves

In [28]python · cell 44

python

if not SKIP_TRAINING and embedding_losses:
    fig, axes = plt.subplots(1, 3, figsize=(14, 4))

    axes[0].plot(embedding_losses, color=COLORS["blue"], linewidth=1.5)
    axes[0].set_title("Phase 1: Embedding")
    axes[0].set_xlabel("Step (×1000)")
    axes[0].set_ylabel("Loss")

    axes[1].plot(supervisor_losses, color=COLORS["blue"], linewidth=1.5)
    axes[1].set_title("Phase 2: Supervisor")
    axes[1].set_xlabel("Step (×1000)")

    axes[2].plot(g_losses, label="Generator", color=COLORS["blue"], linewidth=1.5)
    axes[2].plot(d_losses, label="Discriminator", color=COLORS["amber"], linewidth=1.5)
    axes[2].set_title("Phase 3: Joint")
    axes[2].set_xlabel("Step (×1000)")
    axes[2].legend()

    fig.suptitle("TimeGAN Training Progress", fontsize=14, fontweight="semibold")
    plt.tight_layout()
    plt.show()

Output

/tmp/ipykernel_70/2034406625.py:20: UserWarning: The figure layout has changed to tight
  plt.tight_layout()

<Figure size 1400x400 with 3 Axes>

10. Save Outputs

In [29]python · cell 46

python

CHECKPOINT_DIR.mkdir(parents=True, exist_ok=True)

# Save checkpoint
checkpoint = {
    "embedder": embedder.state_dict(),
    "recovery": recovery.state_dict(),
    "supervisor": supervisor.state_dict(),
    "generator": generator.state_dict(),
    "discriminator": discriminator.state_dict(),
    "config": {
        "tickers": TICKERS,
        "seq_len": SEQ_LEN,
        "hidden_dim": HIDDEN_DIM,
        "num_layers": NUM_LAYERS,
        "train_steps": TRAIN_STEPS,
    },
}
torch.save(checkpoint, CHECKPOINT_PATH)

# Save scaler for denormalization
joblib.dump(scaler, CHECKPOINT_DIR / "scaler.pkl")

Output

['05_synthetic_data/output/timegan/checkpoints/timegan/multi_stock/scaler.pkl']

In [30]python · cell 47

python

# Save metadata
metadata = {
    "generator": "timegan",
    "paper": "Yoon et al., NeurIPS 2019",
    "created_at": datetime.now(UTC).isoformat(),
    "data_format": "multi_stock_adj_close",
    "tickers": TICKERS,
    "config": {
        "seq_len": SEQ_LEN,
        "hidden_dim": HIDDEN_DIM,
        "num_layers": NUM_LAYERS,
        "train_steps": TRAIN_STEPS,
    },
    "evaluation": {
        "discriminative_accuracy": float(disc_accuracy),
        "tstr_ratio": float(tstr_ratio),
    },
}
with open(CHECKPOINT_DIR / "metadata.json", "w") as f:
    json.dump(metadata, f, indent=2)

# Save samples
np.save(CHECKPOINT_DIR / "synthetic.npy", synthetic)
np.save(CHECKPOINT_DIR / "real_train.npy", sequences)
np.save(CHECKPOINT_DIR / "real_holdout.npy", holdout_sequences)

In [31]python · cell 48

python

# Persist PCA + t-SNE 2D coordinates so the book-repo Hard Rule 15 script
# (figures/scripts/generate_figure_5_04_timegan_fidelity.py) can re-render
# the publication figure without retraining. Computed here with the same
# parameters plot_fidelity_comparison() uses internally so the inline render
# and the persisted arrays describe the same projection — including the
# legacy global-RNG seeding (np.random.seed + np.random.choice) so the
# subsample indices match the helper's MT19937 sequence rather than the
# PCG64 sequence np.random.default_rng would emit.
from sklearn.decomposition import PCA
from sklearn.manifold import TSNE

_n_viz = min(1000, len(sequences), len(synthetic))
np.random.seed(42)
_idx_real = np.random.choice(len(sequences), _n_viz, replace=False)
_idx_synth = np.random.choice(len(synthetic), _n_viz, replace=False)
_real_flat = sequences[_idx_real].mean(axis=1)
_synth_flat = synthetic[_idx_synth].mean(axis=1)

_pca = PCA(n_components=min(2, _real_flat.shape[1], _n_viz))
_pca.fit(_real_flat)
_real_pca = _pca.transform(_real_flat)
_synth_pca = _pca.transform(_synth_flat)

_combined = np.vstack([_real_flat, _synth_flat])
_tsne = TSNE(
    n_components=min(2, _real_flat.shape[1]),
    perplexity=min(40, max(2, _n_viz // 4)),
    max_iter=1000,
    random_state=42,
)
_combined_tsne = _tsne.fit_transform(_combined)
_real_tsne = _combined_tsne[:_n_viz]
_synth_tsne = _combined_tsne[_n_viz:]

np.save(CHECKPOINT_DIR / "fidelity_real_pca.npy", _real_pca)
np.save(CHECKPOINT_DIR / "fidelity_synth_pca.npy", _synth_pca)
np.save(CHECKPOINT_DIR / "fidelity_real_tsne.npy", _real_tsne)
np.save(CHECKPOINT_DIR / "fidelity_synth_tsne.npy", _synth_tsne)

print(f"\nSaved to {CHECKPOINT_DIR}/")

Output


Saved to 05_synthetic_data/output/timegan/checkpoints/timegan/multi_stock/

Summary

This notebook implemented TimeGAN (Yoon et al., NeurIPS 2019):

Data: 6 diverse stocks (BA, CAT, DIS, GE, IBM, KO) with adjusted close
Architecture: Five-component system with ModuleList GRUs (matching TF)
Training: Three-phase, step-based approach (10,000 steps per phase)
Evaluation: LSTM-based discriminative and predictive scores

Key Finding

On the six-stock close-price panel used here, the post-training discriminative accuracy is 67.9% (Yoon et al.'s target is ~50%, where the discriminator cannot tell real from synthetic) and the TSTR/TRTR MAE ratio is 1.76 (target ~1.0). Synthetic sequences match the real distribution moments closely (mean 0.376 vs 0.366, std 0.253 vs 0.250) but the discriminator and TSTR diagnostics show that temporal structure is only partially preserved on this run.

Limitations

TimeGAN focuses on matching overall distribution, not tail risk. For alternatives:

Tail risk: See 02_tailgan_tail_risk
Path signatures: See 03_sigcwgan_signatures
Diffusion models: See 05_diffusion_ts