Skip to content

Latest commit

 

History

History
23 lines (18 loc) · 971 Bytes

readme.md

File metadata and controls

23 lines (18 loc) · 971 Bytes

本地加载chatglm2模型

  • 安装git lfs: git-lfs
  • 执行 GIT_LFS_SKIP_SMUDGE=1 git clone https://huggingface.co/THUDM/chatglm2-6b
  • 进入chatglm2-6b文件夹执行git lfs install 和 git lfs pull下载checkpoint
  • 将加载代码中的 THUDM/chatglm2-6b 替换为你本地的 chatglm2-6b 文件夹的路径，即可从本地加载模型。
import os
os.environ['CUDA_VISIBLE_DEVICES']='0'

from transformers import AutoTokenizer, AutoModel

tokenizer = AutoTokenizer.from_pretrained("chatglm2-6b", trust_remote_code=True)
model = AutoModel.from_pretrained("chatglm2-6b", trust_remote_code=True, device='cuda')
model = model.eval()

response, history = model.chat(tokenizer, "你好", history=[])
print(response)
print(history)

response, history = model.chat(tokenizer, "晚上睡不着应该怎么办", history=history)
print(response)