Update app.py

app.py

@@ -47,7 +47,7 @@ class HfModelWrapper:
         input_text = self.build_prompt(user_input, history)
         input_ids = self.tokenizer.encode(input_text, return_tensors="pt").to("cuda")
 
-        streamer = TextIteratorStreamer(self.tokenizer, skip_prompt=True)
+        streamer = TextIteratorStreamer(self.tokenizer, skip_prompt=True, skip_special_tokens=True)
 
         gen_kwargs = {
             "inputs": input_ids,
@@ -509,15 +509,12 @@ def send_message(message, history):
     partial = ""
     idx = 0
     for t in response_generator:
-        if idx <= 5:
+        if idx <= 3:
             idx += 1
             continue
         else:
-            if t != "<|eot_id|>":
-                partial += t
-                yield history + [{"role": "assistant", "content": partial}]    
-            else:
-                continue
+            partial += t
+            yield history + [{"role": "assistant", "content": partial}]    
     
 
 # Create the dashboard