Create app.py

app.py

@@ -0,0 +1,138 @@
+import gradio as gr
+import torch
+from PIL import Image
+from transformers import Qwen3VLForConditionalGeneration, AutoProcessor
+from playwright.sync_api import sync_playwright
+import os
+import time
+
+# --- Configuration ---
+MODEL_PATH = "Qwen/Qwen3-VL-2B-Instruct"
+CPU_DEVICE = "cpu"
+
+# --- Model and Processor Loading ---
+print("Loading model and processor...")
+processor = AutoProcessor.from_pretrained(MODEL_PATH, trust_remote_code=True)
+model = Qwen3VLForConditionalGeneration.from_pretrained(
+    MODEL_PATH,
+    trust_remote_code=True,
+    dtype="auto",
+    device_map="auto"
+)
+print("Model and processor loaded successfully.")
+
+# --- Playwright Screenshot Function ---
+def take_screenshot(url):
+    """
+    Takes a screenshot of a webpage at the given URL using Playwright.
+    """
+    if not url.startswith('http://') and not url.startswith('https://'):
+        url = 'http://' + url
+    try:
+        with sync_playwright() as p:
+            browser = p.chromium.launch(headless=True)
+            page = browser.new_page()
+            page.goto(url, wait_until='networkidle')
+            screenshot_path = f"screenshot_{int(time.time())}.png"
+            page.screenshot(path=screenshot_path, full_page=True)
+            browser.close()
+            return screenshot_path
+    except Exception as e:
+        return f"Error taking screenshot: {str(e)}"
+
+# --- Inference Function ---
+def process_and_generate(image_input, text_prompt):
+    """
+    Processes the image and text prompt, and generates a response from the model.
+    """
+    if image_input is None or not text_prompt.strip():
+        return "Please provide both an image and a text prompt."
+
+    pil_image = Image.fromarray(image_input)
+
+    messages = [
+        {
+            "role": "user",
+            "content": [
+                {"type": "image", "image": pil_image},
+                {"type": "text", "text": text_prompt},
+            ],
+        }
+    ]
+
+    print("Processing inputs and generating response...")
+    try:
+        inputs = processor.apply_chat_template(
+            messages,
+            tokenize=True,
+            add_generation_prompt=True,
+            return_dict=True,
+            return_tensors="pt"
+        )
+        inputs = inputs.to(model.device)
+
+        generated_ids = model.generate(**inputs, max_new_tokens=1024)
+
+        generated_ids_trimmed = [
+            out_ids[len(in_ids):] for in_ids, out_ids in zip(inputs.input_ids, generated_ids)
+        ]
+
+        output_text = processor.batch_decode(
+            generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False
+        )
+
+        return output_text[0]
+
+    except Exception as e:
+        return f"An error occurred during generation: {str(e)}"
+
+# --- Gradio Interface ---
+with gr.Blocks() as demo:
+    gr.Markdown(
+        """
+        # Screenshot to Markdown with Qwen3-VL
+        Enter a URL to take a screenshot, then provide a prompt to generate a markdown document from the image.
+        **Warning:** Running this on a free CPU Space can be slow.
+        """
+    )
+
+    with gr.Row():
+        url_input = gr.Textbox(label="Website URL", placeholder="e.g., www.google.com")
+        screenshot_button = gr.Button("Capture Screenshot")
+
+    with gr.Row():
+        with gr.Column():
+            image_output = gr.Image(type="numpy", label="Screenshot")
+        with gr.Column():
+            text_prompt = gr.Textbox(label="Prompt", placeholder="e.g., Describe this webpage in detail as a markdown document.")
+            submit_button = gr.Button("Generate Markdown")
+
+    with gr.Row():
+        output_text = gr.Markdown(label="Model Output")
+
+
+    def update_image(url):
+        screenshot_path = take_screenshot(url)
+        if isinstance(screenshot_path, str) and os.path.exists(screenshot_path):
+            return screenshot_path
+        else:
+            raise gr.Error(screenshot_path)
+
+    screenshot_button.click(
+        fn=update_image,
+        inputs=url_input,
+        outputs=image_output
+    )
+
+    submit_button.click(
+        fn=process_and_generate,
+        inputs=[image_output, text_prompt],
+        outputs=output_text
+    )
+
+if __name__ == "__main__":
+    # Install playwright browsers
+    import subprocess
+    subprocess.run(["playwright", "install"], check=True)
+    subprocess.run(["playwright", "install-deps"], check=True)
+    demo.launch()