Update app.py

app.py

@@ -9,7 +9,7 @@ import torch
 from PIL import Image
 import gradio as gr
 import spaces
-from transformers import AutoModelForCausalLM, AutoProcessor,TextIteratorStreamer
+from transformers import AutoModelForCausalLM, AutoProcessor, TextIteratorStreamer
 import os
 import time
 from huggingface_hub import hf_hub_download
@@ -44,7 +44,6 @@ model = AutoModelForCausalLM.from_pretrained(
     trust_remote_code=True
 ).to(0)
 processor = AutoProcessor.from_pretrained(MODEL_ID, trust_remote_code=True)
-
 eos_token_id=processor.tokenizer.eos_token_id
 
 
@@ -53,19 +52,19 @@ eos_token_id=processor.tokenizer.eos_token_id
 @spaces.GPU(duration=120, queue=False)
 def stream_chat(message, history: list, system: str, temperature: float, max_new_tokens: int):
     print(message)
-    conversation = [{"role": "system", "content": system or DEFAULT_SYSTEM}]
+    conversation = [{"role": "user", "content": system or DEFAULT_SYSTEM}]
     for prompt, answer in history:
         conversation.extend([{"role": "user", "content": f"<|image_1|>\n{prompt}"}, {"role": "assistant", "content": answer}])
     conversation.append({"role": "user", "content": message['text']})
     
     if message["files"]:
-        image = Image.open(message["files"][0]).convert('RGB')
+        image = Image.open(message["files"][-1]).convert('RGB')
     else:
         image = None
         
     prompt = processor.tokenizer.apply_chat_template(conversation, tokenize=False, add_generation_prompt=True)
-    inputs = processor(prompt, [image], return_tensors="pt").to(0) 
-
+    inputs = processor(prompt, images=[image], return_tensors="pt").to(0)
+    
     generate_kwargs = dict(
         max_new_tokens=max_new_tokens,
         temperature=temperature,