Spaces:
Sleeping
Sleeping
| import os | |
| import gradio as gr | |
| import requests | |
| import pandas as pd | |
| from transformers import pipeline | |
| # --- Constants --- | |
| DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space" | |
| HF_MODEL_NAME = "facebook/bart-large-mnli" # Free model that works in Spaces | |
| # --- Agent Definition --- | |
| class BasicAgent: | |
| def __init__(self): | |
| print("Initializing Agent...") | |
| try: | |
| self.llm = pipeline( | |
| "text-generation", | |
| model=HF_MODEL_NAME, | |
| device_map="auto" | |
| ) | |
| except Exception as e: | |
| print(f"LLM initialization failed: {e}") | |
| self.llm = None | |
| def __call__(self, question: str) -> str: | |
| if not self.llm: | |
| return "Default answer (LLM not available)" | |
| try: | |
| response = self.llm(question, max_length=100) | |
| return response[0]['generated_text'] | |
| except Exception as e: | |
| return f"Error: {str(e)}" | |
| def run_and_submit_all(): | |
| """Simplified version that works with Gradio auth""" | |
| # Get username from Gradio's auth system | |
| username = os.getenv("GRADIO_AUTH_USERNAME") | |
| if not username: | |
| return "Please login first", None | |
| space_id = os.getenv("SPACE_ID") | |
| api_url = DEFAULT_API_URL | |
| agent = BasicAgent() | |
| # Fetch questions | |
| try: | |
| response = requests.get(f"{api_url}/questions", timeout=15) | |
| questions = response.json() | |
| except Exception as e: | |
| return f"Failed to get questions: {str(e)}", None | |
| # Process questions | |
| results = [] | |
| answers = [] | |
| for q in questions: | |
| try: | |
| answer = agent(q.get("question", "")) | |
| answers.append({ | |
| "task_id": q.get("task_id"), | |
| "submitted_answer": answer | |
| }) | |
| results.append({ | |
| "Task ID": q.get("task_id"), | |
| "Question": q.get("question"), | |
| "Answer": answer | |
| }) | |
| except Exception as e: | |
| results.append({ | |
| "Task ID": q.get("task_id"), | |
| "Question": q.get("question"), | |
| "Answer": f"Error: {str(e)}" | |
| }) | |
| # Submit answers | |
| try: | |
| response = requests.post( | |
| f"{api_url}/submit", | |
| json={ | |
| "username": username, | |
| "agent_code": f"https://huggingface.co/spaces/{space_id}", | |
| "answers": answers | |
| }, | |
| timeout=60 | |
| ) | |
| result = response.json() | |
| return ( | |
| f"Success! Score: {result.get('score', 'N/A')}%\n" | |
| f"Correct: {result.get('correct_count', 0)}/{result.get('total_attempted', 0)}", | |
| pd.DataFrame(results) | |
| ) | |
| except Exception as e: | |
| return f"Submission failed: {str(e)}", pd.DataFrame(results) | |
| # --- Gradio Interface --- | |
| with gr.Blocks() as demo: | |
| gr.Markdown("# LLM Agent Evaluation") | |
| with gr.Accordion("Instructions", open=False): | |
| gr.Markdown(""" | |
| 1. Click the login button | |
| 2. Authorize with your Hugging Face account | |
| 3. Click 'Run Evaluation' | |
| """) | |
| gr.LoginButton() | |
| run_btn = gr.Button("Run Evaluation", variant="primary") | |
| status = gr.Textbox(label="Status") | |
| results = gr.DataFrame(label="Results", wrap=True) | |
| run_btn.click( | |
| fn=run_and_submit_all, | |
| outputs=[status, results] | |
| ) | |
| if __name__ == "__main__": | |
| demo.launch(auth_message="Please login with your Hugging Face account") |