Spaces:
Sleeping
Sleeping
| import gradio as gr | |
| import requests | |
| import base64 | |
| from PIL import Image | |
| from io import BytesIO | |
| print("=== DEBUG: Starting app.py ===") | |
| def encode_image_to_base64(image: Image.Image) -> str: | |
| buffered = BytesIO() | |
| image.save(buffered, format="JPEG") | |
| img_str = base64.b64encode(buffered.getvalue()).decode() | |
| return f"data:image/jpeg;base64,{img_str}" | |
| def query_vllm_api(image, temperature, max_tokens=12_000): | |
| print(f"=== DEBUG: query_vllm_api called with image={image is not None}, temp={temperature} ===") | |
| if image is None: | |
| return "No image provided", "No image provided", "Please upload an image first." | |
| try: | |
| messages = [] | |
| # Optional: Resize image if needed (to avoid huge uploads) | |
| max_size = 1024 | |
| if max(image.size) > max_size: | |
| ratio = max_size / max(image.size) | |
| new_size = tuple(int(dim * ratio) for dim in image.size) | |
| image = image.resize(new_size, Image.Resampling.LANCZOS) | |
| image_b64 = encode_image_to_base64(image) | |
| messages.append({ | |
| "role": "user", | |
| "content": [ | |
| {"type": "image_url", "image_url": {"url": image_b64}} | |
| ] | |
| }) | |
| payload = { | |
| "model": "numind/NuMarkdown-8B-Thinking", | |
| "messages": messages, | |
| "max_tokens": max_tokens, | |
| "temperature": temperature | |
| } | |
| print("=== DEBUG: About to make vLLM API request ===") | |
| response = requests.post( | |
| "http://localhost:8000/v1/chat/completions", | |
| json=payload, | |
| timeout=60 | |
| ) | |
| response.raise_for_status() | |
| data = response.json() | |
| result = data["choices"][0]["message"]["content"] | |
| # Handle the thinking/answer parsing | |
| try: | |
| reasoning = result.split("<think>")[1].split("</think>")[0] | |
| answer = result.split("<answer>")[1].split("</answer>")[0] | |
| except IndexError: | |
| # If no thinking tags, return the full result | |
| reasoning = "No thinking trace found" | |
| answer = result | |
| return reasoning, answer, answer | |
| except requests.exceptions.RequestException as e: | |
| error_msg = f"API request failed: {e}" | |
| print(f"=== DEBUG: Request error: {error_msg} ===") | |
| return error_msg, error_msg, error_msg | |
| except Exception as e: | |
| error_msg = f"Unexpected error: {e}" | |
| print(f"=== DEBUG: Unexpected error: {error_msg} ===") | |
| return error_msg, error_msg, error_msg | |
| print("=== DEBUG: Creating Gradio interface ===") | |
| with gr.Blocks(title="NuMarkdown-8B-Thinking", theme=gr.themes.Soft()) as demo: | |
| gr.HTML(""" | |
| <div style="text-align: center; padding: 20px; background: linear-gradient(90deg, #667eea 0%, #764ba2 100%); border-radius: 10px; margin-bottom: 20px;"> | |
| <h1 style="color: white; margin: 0; font-size: 2.5em; font-weight: bold;">👁️ NuMarkdown-8B-Thinking</h1> | |
| <p style="color: rgba(255,255,255,0.9); margin: 10px 0; font-size: 1.2em;">Upload an image to convert to Markdown!</p> | |
| </div> | |
| """) | |
| with gr.Row(): | |
| with gr.Column(): | |
| temperature = gr.Slider(0.1, 1.5, value=0.6, step=0.1, label="Temperature") | |
| img_in = gr.Image(type="pil", label="Upload Image") | |
| btn = gr.Button("Generate Response") | |
| with gr.Column(): | |
| thinking = gr.Textbox(label="Thinking Trace", lines=10) | |
| raw_answer = gr.Textbox(label="Raw Output", lines=5) | |
| output = gr.Markdown(label="Response") | |
| btn.click( | |
| query_vllm_api, | |
| inputs=[img_in, temperature], | |
| outputs=[thinking, raw_answer, output], | |
| ) | |
| print("=== DEBUG: Gradio interface created ===") | |
| if __name__ == "__main__": | |
| print("=== DEBUG: About to launch Gradio ===") | |
| demo.launch( | |
| server_name="0.0.0.0", | |
| server_port=7860, | |
| share=False | |
| ) | |
| print("=== DEBUG: Gradio launched ===") |