Update app.py

app.py

@@ -1,12 +1,11 @@
 import streamlit as st
-from transformers import AutoProcessor, Qwen2VLForConditionalGeneration, AutoConfig
+from transformers import AutoProcessor, Qwen2VLForConditionalGeneration
 from PIL import Image
 import torch
 
-# Load the processor and model configuration
+# Load the processor and model directly
 processor = AutoProcessor.from_pretrained("Qwen/Qwen2-VL-2B-Instruct")
-config = AutoConfig.from_pretrained("Qwen/Qwen2-VL-2B-Instruct")
-model = Qwen2VLForConditionalGeneration.from_pretrained("Qwen/Qwen2-VL-2B-Instruct", config=config)
+model = Qwen2VLForConditionalGeneration.from_pretrained("Qwen/Qwen2-VL-2B-Instruct")
 
 # Streamlit app
 st.title("Image Description Generator")
@@ -37,10 +36,6 @@ if uploaded_file is not None:
         messages, tokenize=False, add_generation_prompt=True
     )
 
-    # Debugging: Display the generated text
-    st.write("Generated text for processing:")
-    st.write(text)
-
     # Pass the image to the processor
     inputs = processor(
         text=[text],
@@ -50,10 +45,6 @@ if uploaded_file is not None:
     )
     inputs = inputs.to("cpu")
 
-    # Debugging: Display the inputs
-    st.write("Inputs for the model:")
-    st.write(inputs)
-
     # Inference: Generation of the output
     generated_ids = model.generate(**inputs, max_new_tokens=128)
     generated_ids_trimmed = [
@@ -63,9 +54,5 @@ if uploaded_file is not None:
         generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False
     )
     
-    # Debugging: Display the raw output text
-    st.write("Raw output text:")
-    st.write(output_text)
-
     st.write("Description:")
-    st.write(output_text[0])
+    st.write(output_text)