Re-enable TOKENIZERS_PARALLELISM

app.py

@@ -19,12 +19,18 @@ processor: Qwen2VLProcessor = AutoProcessor.from_pretrained(model_path, min_pixe
 
 @GPU
 def infer(
-    inputs: BatchFeature,
+    inputs: tuple,
     max_tokens: int,
     temperature: float,
     top_p: float,
 ):
-    inputs = inputs.to("cuda")
+    inputs = processor(
+        text=[inputs[0]],
+        images=inputs[1],
+        videos=inputs[2],
+        padding=True,
+        return_tensors="pt",
+    ).to("cuda")
     streamer = TextIteratorStreamer(processor, skip_prompt=True, skip_special_tokens=True)
     kwargs = dict(
         **inputs,
@@ -53,14 +59,7 @@ def respond(
     messages.append({"role": "user", "content": message})
     text_inputs = processor.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
     image_inputs, video_inputs = process_vision_info(messages)
-    inputs = processor(
-        text = [text_inputs],
-        images = image_inputs,
-        videos = video_inputs,
-        padding = True,
-        return_tensors = "pt",
-    )
-    for response in infer(inputs, max_tokens, temperature, top_p):
+    for response in infer((text_inputs, image_inputs, video_inputs), max_tokens, temperature, top_p):
         yield response
 
 app = ChatInterface(