Update multipurpose_chatbot/engines/transformers_engine.py

multipurpose_chatbot/engines/transformers_engine.py

@@ -546,7 +546,7 @@ class TransformersEngine(BaseEngine):
                         raise gr.Error(message_safety)
 
     # @maybe_spaces_gpu
-    def generate_yield_string(self, prompt, temperature, max_tokens, stop_strings: Optional[Tuple[str]] = None, **kwargs):
+    def generate_yield_string(self, prompt, temperature=0.7, max_tokens=1024, stop_strings: Optional[Tuple[str]] = None, **kwargs):
         
         # ! MUST PUT INSIDE torch.no_grad() otherwise it will overflow OOM
         import sys
@@ -564,7 +564,7 @@ class TransformersEngine(BaseEngine):
                 inputs = self.tokenizer(prompt, return_tensors='pt')
                 # whether to print the full prompts
                 retok_full_prompt = self.tokenizer.decode(inputs.input_ids[0], skip_special_tokens=False)
-                print(f"retok_full_prompt:\n{retok_full_prompt}>>>>")
+                # print(f"retok_full_prompt:\n{retok_full_prompt}>>>>")
                 begin_bos = inputs.input_ids[0][0] == self.tokenizer.bos_token_id
                 print(f'begin_bos: {begin_bos}')