Update app.py

app.py

@@ -5,7 +5,7 @@ import requests
 import spaces
 from transformers import AutoProcessor, Idefics3ForConditionalGeneration, TextIteratorStreamer, StoppingCriteria, StoppingCriteriaList
 
-base_model_id = "Andres77872/SmolVLM-500M-anime-caption-v0.1"
+base_model_id = "Andres77872/SmolVLM-500M-anime-caption-v0.2"
 
 processor = AutoProcessor.from_pretrained(base_model_id)
 model = Idefics3ForConditionalGeneration.from_pretrained(
@@ -50,7 +50,7 @@ def caption_anime_image_stream(image):
     inputs = processor(text=[prompt], images=[[image]], return_tensors='pt', padding=True, size=size)
     inputs = {k: v.to(model.device) for k, v in inputs.items()}
     
-    stop_sequence = "</QUERY>"
+    stop_sequence = "</RATING>"
     streamer = TextIteratorStreamer(
         processor.tokenizer,
         skip_prompt=True,
@@ -64,7 +64,7 @@ def caption_anime_image_stream(image):
             **inputs,
             streamer=streamer,
             do_sample=False,
-            max_new_tokens=512,
+            max_new_tokens=1024,
             pad_token_id=processor.tokenizer.pad_token_id,
             stopping_criteria=custom_stopping_criteria,
         )
@@ -81,7 +81,7 @@ demo = gr.Interface(
     caption_anime_image_stream,
     inputs=gr.Image(type="pil", label="Anime Image"),
     outputs=gr.Textbox(lines=8, label="Caption"),
-    title="SmolVLM-500M-Anime-Caption Demo",
+    title="SmolVLM-500M-anime-caption-v0.2 Demo",
     description="Upload an anime-style image to generate a caption.",
     # Enable live streaming:
     allow_flagging="auto",