Update app.py

app.py

@@ -16,18 +16,18 @@ HF_TOKEN = os.environ.get("HF_TOKEN", None)
 model_id = 'mistralai/Mistral-7B-Instruct-v0.3'
 client = InferenceClient(model_id)
 
-# Define stopping criteria
-class StopOnTokens:
-    def __call__(self, input_ids, scores, **kwargs):
-        for stop_ids in stop_token_ids:
-            if torch.eq(input_ids[0][-len(stop_ids):], stop_ids).all():
-                return True
-        return False
-
-# Define stopping criteria list
-stop_list = ['\nHuman:', '\n```\n']
-stop_token_ids = [client.tokenizer(x)['input_ids'] for x in stop_list]
-stop_token_ids = [torch.LongTensor(x).to(cuda.current_device() if cuda.is_available() else 'cpu') for x in stop_token_ids]
+# # Define stopping criteria
+# class StopOnTokens:
+#     def __call__(self, input_ids, scores, **kwargs):
+#         for stop_ids in stop_token_ids:
+#             if torch.eq(input_ids[0][-len(stop_ids):], stop_ids).all():
+#                 return True
+#         return False
+
+# # Define stopping criteria list
+# stop_list = ['\nHuman:', '\n```\n']
+# stop_token_ids = [client.tokenizer(x)['input_ids'] for x in stop_list]
+# stop_token_ids = [torch.LongTensor(x).to(cuda.current_device() if cuda.is_available() else 'cpu') for x in stop_token_ids]
 
 # Create text generation pipeline
 def generate(prompt, history, system_prompt=None, temperature=0.2, max_new_tokens=1024, top_p=0.95, repetition_penalty=1.0):