Update app.py

app.py

@@ -1,8 +1,8 @@
-# app.py
 import gradio as gr
 from transformers import AutoModelForVision2Seq, AutoProcessor
 import torch
 from PIL import Image
+import os
 
 # Load Qwen-VL model and processor
 model_id = "Qwen/Qwen-VL-Chat"
@@ -11,6 +11,9 @@ model = AutoModelForVision2Seq.from_pretrained(model_id, torch_dtype=torch.float
 
 # Inference function
 def ocr_with_qwen(image):
+    if image is None:
+        image = Image.open("test.png")
+
     prompt = "<|im_start|>system\nYou are a helpful assistant. Extract all text from the image and output only the text.<|im_end|>\n<|im_start|>user\n"
     inputs = processor(images=image, text=prompt, return_tensors="pt").to(model.device)
     outputs = model.generate(**inputs, max_new_tokens=512)
@@ -20,8 +23,9 @@ def ocr_with_qwen(image):
 # Gradio UI
 gr.Interface(
     fn=ocr_with_qwen,
-    inputs=gr.Image(type="pil", label="Upload Image (test.jpg)"),
+    inputs=gr.Image(type="pil", label="Upload Image (defaults to test.png if none uploaded)", optional=True),
     outputs=gr.Textbox(label="Extracted Text"),
     title="OCR with Qwen2.5-VL",
-    description="Upload an image to extract text using Qwen-VL model."
+    description="Upload an image to extract text using Qwen-VL model. If no image is uploaded, test.png is used.",
+    examples=[["test.png"]]
 ).launch()