fix: small fix

app.py

@@ -13,17 +13,23 @@ def greet_json():
 
 @app.post("/")
 async def aria_image_to_text(request: Request):
+    print(1)
     data = await request.json()
+    print(2)
     image_url = data.get("image_url")
+    print(3)
+    print('image_url')
+    print(image_url)
     image = Image.open(requests.get(image_url, stream=True).raw)
-
+    print(4)
     model_id_or_path = "rhymes-ai/Aria"
+    print(5)
     model = AriaForConditionalGeneration.from_pretrained(
         model_id_or_path, device_map="auto", torch_dtype=torch.bfloat16
     )
-
+    print(6)
     processor = AriaProcessor.from_pretrained(model_id_or_path)
-
+    print(7)
     messages = [
         {
             "role": "user",
@@ -34,11 +40,15 @@ async def aria_image_to_text(request: Request):
         }
     ]
 
+    print(8)
     text = processor.apply_chat_template(messages, add_generation_prompt=True)
+    print(9)
     inputs = processor(text=text, images=image, return_tensors="pt")
+    print(10)
     inputs['pixel_values'] = inputs['pixel_values'].to(torch.bfloat16)
+    print(11)
     inputs.to(model.device)
-
+    print(12)
     output = model.generate(
         **inputs,
         max_new_tokens=15,
@@ -47,21 +57,26 @@ async def aria_image_to_text(request: Request):
         do_sample=True,
         temperature=0.9,
     )
+    print(13)
     output_ids = output[0][inputs["input_ids"].shape[1]:]
+    print(14)
     response = processor.decode(output_ids, skip_special_tokens=True)
+    print(15)
     return {"response": response}
 
 @app.get("/aria-test")
 def aria_test():
+    print(1)
     model_id_or_path = "rhymes-ai/Aria"
+    print(2)
     model = AriaForConditionalGeneration.from_pretrained(
         model_id_or_path, device_map="auto", torch_dtype=torch.bfloat16
     )
-
+    print(3)
     processor = AriaProcessor.from_pretrained(model_id_or_path)
-
+    print(4)
     image = Image.open(requests.get("http://images.cocodataset.org/val2017/000000039769.jpg", stream=True).raw)
-
+    print(5)
     messages = [
         {
             "role": "user",
@@ -71,12 +86,15 @@ def aria_test():
             ],
         }
     ]
-
+    print(6)    
     text = processor.apply_chat_template(messages, add_generation_prompt=True)
+    print(7)
     inputs = processor(text=text, images=image, return_tensors="pt")
+    print(8)
     inputs['pixel_values'] = inputs['pixel_values'].to(torch.bfloat16)
+    print(9)
     inputs.to(model.device)
-
+    print(10)
     output = model.generate(
         **inputs,
         max_new_tokens=15,
@@ -86,5 +104,7 @@ def aria_test():
         temperature=0.9,
     )
     output_ids = output[0][inputs["input_ids"].shape[1]:]
+    print(11)
     response = processor.decode(output_ids, skip_special_tokens=True)
-    return {"response": response}
+    print(12)
+    return {"response": response}