Spaces:
Sleeping
Sleeping
| # app.py for Hugging Face Spaces | |
| import os | |
| import requests | |
| import gradio as gr | |
| import time | |
| from datetime import datetime | |
| # Configuration abdelac/Mistral_Test | |
| HF_TOKEN = os.getenv("HF_TOKEN", "") | |
| MODEL_NAME = os.getenv("MODEL_NAME", "abdelac/Mistral_Test") | |
| # Use different API URL format for Spaces | |
| API_URL = f"https://router.huggingface.co/models/{MODEL_NAME}" | |
| HEADERS = {"Authorization": f"Bearer {HF_TOKEN}"} if HF_TOKEN else {} | |
| # Cache for API health | |
| api_status_cache = {"last_check": 0, "status": None} | |
| def check_api_status(): | |
| """Check if the API is accessible""" | |
| try: | |
| response = requests.head(API_URL, headers=HEADERS, timeout=5) | |
| return { | |
| "available": response.status_code in [200, 503], | |
| "status_code": response.status_code, | |
| "message": "API is accessible" if response.status_code == 200 else "Model is loading" | |
| } | |
| except: | |
| return {"available": False, "status_code": None, "message": "Cannot connect to API"} | |
| def query_model(prompt, max_tokens=256, temperature=0.7): | |
| """Query the model with error handling""" | |
| if not prompt.strip(): | |
| return "⚠️ Please enter a prompt" | |
| if not HF_TOKEN: | |
| return "🔐 Please add your HF_TOKEN in Space Settings → Repository secrets" | |
| payload = { | |
| "inputs": prompt, | |
| "parameters": { | |
| "max_new_tokens": max_tokens, | |
| "temperature": temperature, | |
| "return_full_text": False | |
| }, | |
| "options": {"wait_for_model": True} | |
| } | |
| try: | |
| start = time.time() | |
| response = requests.post(API_URL, headers=HEADERS, json=payload, timeout=30) | |
| if response.status_code == 200: | |
| result = response.json() | |
| elapsed = time.time() - start | |
| if isinstance(result, list) and len(result) > 0: | |
| text = result[0].get("generated_text", str(result)) | |
| return f"{text}\n\n⏱️ Generated in {elapsed:.2f}s" | |
| else: | |
| return f"Response format unexpected: {result}" | |
| elif response.status_code == 503: | |
| return "🔄 Model is loading. Please wait 30 seconds and try again." | |
| elif response.status_code == 401: | |
| return "🔐 Invalid token. Please check your HF_TOKEN." | |
| else: | |
| error = response.json().get("error", response.text[:200]) | |
| return f"❌ Error {response.status_code}: {error}" | |
| except requests.exceptions.Timeout: | |
| return "⏱️ Request timeout. Try reducing max tokens." | |
| except Exception as e: | |
| return f"⚠️ Error: {str(e)}" | |
| # Create Gradio Interface | |
| with gr.Blocks( | |
| title="Mistral Test Model", | |
| theme=gr.themes.Soft(), | |
| css=""" | |
| .gradio-container {max-width: 800px; margin: auto;} | |
| .status {padding: 10px; border-radius: 5px; margin: 10px 0;} | |
| .ok {background: #d4edda; color: #155724;} | |
| .warn {background: #fff3cd; color: #856404;} | |
| .error {background: #f8d7da; color: #721c24;} | |
| """ | |
| ) as demo: | |
| gr.Markdown(""" | |
| # 🤖 Mistral Test Model | |
| ### Testing Hugging Face Model Deployment | |
| This Space demonstrates deployment of the `abdelac/Mistral_Test` model. | |
| """) | |
| # Status display | |
| status_display = gr.Markdown("") | |
| # Main interface | |
| with gr.Row(): | |
| with gr.Column(): | |
| prompt = gr.Textbox( | |
| label="Your Prompt", | |
| placeholder="Type your message here...", | |
| lines=5 | |
| ) | |
| with gr.Row(): | |
| max_tokens = gr.Slider( | |
| minimum=32, | |
| maximum=512, | |
| value=256, | |
| step=32, | |
| label="Max Tokens" | |
| ) | |
| temperature = gr.Slider( | |
| minimum=0.1, | |
| maximum=1.5, | |
| value=0.7, | |
| step=0.1, | |
| label="Temperature" | |
| ) | |
| generate_btn = gr.Button("Generate", variant="primary") | |
| clear_btn = gr.Button("Clear") | |
| with gr.Column(): | |
| output = gr.Textbox( | |
| label="Model Response", | |
| lines=8, | |
| interactive=False | |
| ) | |
| # Examples | |
| gr.Examples( | |
| examples=[ | |
| ["Explain quantum computing in simple terms:"], | |
| ["Write a short poem about AI:"], | |
| ["What is the capital of France?"], | |
| ["How to make a cup of coffee:"] | |
| ], | |
| inputs=prompt, | |
| label="Try these examples" | |
| ) | |
| # Instructions | |
| with gr.Accordion("📖 Setup Instructions", open=False): | |
| gr.Markdown(f""" | |
| ## How to Set Up This Space: | |
| 1. **Click "Duplicate this Space"** (top right) to create your own copy | |
| 2. **Add your HF_TOKEN** in Settings → Repository secrets: | |
| - Go to [huggingface.co/settings/tokens](https://huggingface.co/settings/tokens) | |
| - Create a new token with "read" access | |
| - Add it as `HF_TOKEN` in your Space settings | |
| 3. **Optional**: Change model in Settings → Variables: | |
| - Add variable: `MODEL_NAME` = `abdelac/Mistral_Test` | |
| - Or use any other model name | |
| 4. **The Space will automatically deploy** with your configuration | |
| **Current Model**: `{MODEL_NAME}` | |
| """) | |
| # Functions | |
| def update_status(): | |
| status = check_api_status() | |
| if status["available"]: | |
| if status["status_code"] == 200: | |
| return f"""<div class='status ok'>✅ API Status: Ready | |
| <br><small>Model can be queried successfully</small></div>""" | |
| else: | |
| return f"""<div class='status warn'>⚠️ API Status: Loading | |
| <br><small>Model is starting up (Code: {status['status_code']})</small></div>""" | |
| else: | |
| return f"""<div class='status error'>❌ API Status: Unavailable | |
| <br><small>{status['message']}</small></div>""" | |
| def clear(): | |
| return ["", 256, 0.7, ""] | |
| # Event handlers | |
| generate_btn.click( | |
| fn=query_model, | |
| inputs=[prompt, max_tokens, temperature], | |
| outputs=output | |
| ) | |
| clear_btn.click( | |
| fn=clear, | |
| outputs=[prompt, max_tokens, temperature, output] | |
| ) | |
| # Auto-check on load | |
| demo.load( | |
| fn=update_status, | |
| outputs=status_display | |
| ) | |
| # Launch | |
| if __name__ == "__main__": | |
| # Print debug info | |
| print("=" * 50) | |
| print(f"Model: {MODEL_NAME}") | |
| print(f"Token present: {'Yes' if HF_TOKEN else 'No'}") | |
| print(f"API URL: {API_URL}") | |
| print("=" * 50) | |
| demo.launch(share=False) |