Update app.py

app.py

@@ -30,11 +30,26 @@ model = AutoModelForVision2Seq.from_pretrained(model_id, torch_dtype="auto", dev
 
 @spaces.GPU
 def stream_chat(message: Dict[str, str], history: list):
+    # Turn 1:
     # {'text': 'what is this', 'files': ['image-xxx.jpg']}
-    print(message)
-    print(history)
+    # []
+
+    # Turn 2:
+    # {'text': 'continue?', 'files': []}
+    # [[('image-xxx.jpg',), None], ['what is this', 'a image.']]
+
+    if len(message["files"]) != 0:
+        image_path = message["files"][0]
+
+    if isinstance(history[0][0], tuple):
+        image_path = history[0][0][0]
+        history.pop(0)
+
+    if image_path is not None:
+        image = Image.open(image_path)
+    else:
+        image = Image.new("RGB", (100, 100), (255, 255, 255))
 
-    image = Image.open(message["files"][0])
     pixel_values = processor(images=[image], return_tensors="pt").to(model.device)["pixel_values"]
 
     conversation = []