Update inference_fine_tune.py

inference_fine_tune.py

@@ -115,7 +115,7 @@ def generate_response(prompt:str):
     model  = get_model(config, tokenizer.get_vocab_size()).to(device)
     model_path = get_weights_file_path(config,config['preload'])
     model.eval()        
-    state = torch.load(model_path)
+    state = torch.load(model_path,map_location=torch.device('cpu'))
     model.load_state_dict(state['model_state_dict'])
     word = ""
     input_tokens = tokenizer.encode(prompt).ids