Chapter 28
加载基础模型
Notebooktransformers11 cells
In [ ]python · cell 1
python
from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel加载基础模型
In [ ]python · cell 3
python
model = AutoModelForCausalLM.from_pretrained("Langboat/bloom-1b4-zh")
tokenizer = AutoTokenizer.from_pretrained("Langboat/bloom-1b4-zh")加载Lora模型
In [ ]python · cell 5
python
p_model = PeftModel.from_pretrained(model, model_id="./chatbot/checkpoint-500/")
p_modelIn [ ]python · cell 6
python
ipt = tokenizer("Human: {}\n{}".format("考试有哪些技巧？", "").strip() + "\n\nAssistant: ", return_tensors="pt")
tokenizer.decode(p_model.generate(**ipt, do_sample=False)[0], skip_special_tokens=True)模型合并
In [ ]python · cell 8
python
merge_model = p_model.merge_and_unload()
merge_modelIn [ ]python · cell 9
python
ipt = tokenizer("Human: {}\n{}".format("考试有哪些技巧？", "").strip() + "\n\nAssistant: ", return_tensors="pt")
tokenizer.decode(merge_model.generate(**ipt, do_sample=False)[0], skip_special_tokens=True)完整模型保存
In [ ]python · cell 11
python
merge_model.save_pretrained("./chatbot/merge_model")