Update app.py

app.py

@@ -15,7 +15,7 @@ MODEL, PROCESSOR = init_qwen_model()
 
 # Streamlit app title
 st.title("OCR Image Text Extraction")
-
+st.subheader("I used Qwen2-VL-7B-Instruct model to get better accuracy but as it is running on CPU it takes 25-30 minutes to run it. So please have patience.")
 # File uploader for images
 uploaded_file = st.file_uploader("Choose an image...", type=["png", "jpg", "jpeg"])
 
@@ -33,7 +33,7 @@ if uploaded_file is not None:
                         "type": "image",
                         "image": image,
                     },
-                    {"type": "text", "text": "Run Optical Character recognition on the image."},
+                    {"type": "text", "text": "Run Optical Character recognition on the image for Hindi and English."},
                 ],
             }
         ]
@@ -53,7 +53,7 @@ if uploaded_file is not None:
         inputs = inputs.to("cpu")
 
         # Inference: Generation of the output
-        generated_ids = MODEL.generate(**inputs, max_new_tokens=512)
+        generated_ids = MODEL.generate(**inputs, max_new_tokens=256)
         generated_ids_trimmed = [
             out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids)
         ]