Refactor to use upstream automodel.

app.py

@@ -19,7 +19,7 @@ T4级别的GPU已经可以很胜任这个任务了。
 import torch
 import sys
 
-from transformers import AutoTokenizer, GenerationConfig, Automodel
+from transformers import AutoTokenizer, GenerationConfig, AutoModel
 
 model = AutoModel.from_pretrained("THUDM/chatglm-6b", trust_remote_code=True).float()
 tokenizer = AutoTokenizer.from_pretrained("THUDM/chatglm-6b", trust_remote_code=True)