cpu3

app.py

@@ -8,9 +8,6 @@ from accelerate import infer_auto_device_map, init_empty_weights, load_checkpoin
 import gradio as gr
 import time
 
-#使用cpu
-map_location=torch.device('cpu')
-
 model_path = "THUDM/chatglm-6b-int4"
 # 载入Tokenizer
 tokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=True)
@@ -18,7 +15,7 @@ tokenizer = AutoTokenizer.from_pretrained(model_path, trust_remote_code=True)
 config = AutoConfig.from_pretrained(model_path, trust_remote_code=True, pre_seq_len=128)
 model = AutoModel.from_pretrained(model_path, config=config, trust_remote_code=True)
 # 此处使用你的 ptuning 工作目录
-prefix_state_dict = torch.load(os.path.join("./xiaowo", "pytorch_model.bin"))
+prefix_state_dict = torch.load(os.path.join("./xiaowo", "pytorch_model.bin"), map_location=torch.device('cpu'))
 new_prefix_state_dict = {}
 for k, v in prefix_state_dict.items():
     new_prefix_state_dict[k[len("transformer.prefix_encoder."):]] = v