Update app.py

app.py

@@ -69,13 +69,11 @@ def bot_streaming(message, history):
     else:
         conversation.append({"role": "user", "content": message['text']})
     print(f"prompt is -\n{conversation}")
-    #prompt = f"""User:<image>\n{message['text']} Falcon:"""
     prompt = processor.tokenizer.apply_chat_template(conversation, tokenize=False, add_generation_prompt=True)
     image = Image.open(image)
     inputs = processor(prompt, image, return_tensors="pt").to("cuda:0") 
-    #inputs = processor(prompt, image, return_tensors='pt').to(0, torch.float16)
 
-    streamer = TextIteratorStreamer(processor, **{"skip_special_tokens": True, "skip_prompt": True, 'clean_up_tokenization_spaces':False,}) # "eos_token_id":processor.tokenizer.eos_token_id})
+    streamer = TextIteratorStreamer(processor, **{"skip_special_tokens": True, "skip_prompt": True, 'clean_up_tokenization_spaces':False,}) 
     generation_kwargs = dict(inputs, streamer=streamer, max_new_tokens=1024, do_sample=False, temperature=0.0, eos_token_id=processor.tokenizer.eos_token_id,)
 
     thread = Thread(target=model.generate, kwargs=generation_kwargs)
@@ -83,9 +81,6 @@ def bot_streaming(message, history):
 
     buffer = ""
     for new_text in streamer:
-        # find <|eot_id|> and remove it from the new_text
-        #if "<|eot_id|>" in new_text:
-        #    new_text = new_text.split("<|eot_id|>")[0]
         buffer += new_text
         yield buffer