Update infernece_FinGPT.py

infernece_FinGPT.py

@@ -3,14 +3,12 @@ import platform
 import signal
 from transformers import AutoTokenizer, AutoModel
 import readline
-from peft import PeftModel  
 
-base_model = '../FinGPT/THUDM/chatglm2-6b' # 模型地址，可在对应模型官网进行下载模型文件,也可以直接使用huggingface的模型加载
-lora_model = '../FinGPT/saves/FinGPT/pretrained_lora/fingpt-mt_chatglm2-6b_lora'     # 需要下载，本地模型预训练权重,可在 <https://huggingface.co/FinGPT>下载对应模型的预训练权重
+
+base_model = '/home/wanglch/projects/FinGPT/FinGPT_mt_chatglm2-6b-merged'
 
 tokenizer = AutoTokenizer.from_pretrained(base_model, trust_remote_code=True)
 model = AutoModel.from_pretrained(base_model, trust_remote_code=True, device_map = "auto")
-lora_model = PeftModel.from_pretrained(model, lora_model)
 
 
 # 多显卡支持，使用下面两行代替上面一行，将num_gpus改为你实际的显卡数量