use inference_mode

modeling_chatglm.py

@@ -1014,7 +1014,7 @@ class ChatGLMForConditionalGeneration(ChatGLMPreTrainedModel):
         inputs = inputs.to(self.device)
         return inputs
 
-    @torch.no_grad()
+    @torch.inference_mode()
     def chat(self, tokenizer, query: str, history: List[Tuple[str, str]] = None, max_length: int = 8192, num_beams=1,
              do_sample=True, top_p=0.8, temperature=0.8, logits_processor=None, **kwargs):
         if history is None:
@@ -1032,7 +1032,7 @@ class ChatGLMForConditionalGeneration(ChatGLMPreTrainedModel):
         history = history + [(query, response)]
         return response, history
 
-    @torch.no_grad()
+    @torch.inference_mode()
     def stream_chat(self, tokenizer, query: str, history: List[Tuple[str, str]] = None, past_key_values=None,
                     max_length: int = 8192, do_sample=True, top_p=0.8, temperature=0.8, logits_processor=None,
                     return_past_key_values=False, **kwargs):
@@ -1069,7 +1069,7 @@ class ChatGLMForConditionalGeneration(ChatGLMPreTrainedModel):
                 else:
                     yield response, new_history
 
-    @torch.no_grad()
+    @torch.inference_mode()
     def stream_generate(
             self,
             input_ids,