Supastrikas-004/evaluation-framework

1---2title: Agentic Evaluation Framework3emoji: 🤖4colorFrom: indigo5colorTo: blue6sdk: gradio7sdk_version: 5.45.08app_file: app.py9pinned: false10---11 12# Agentic Evaluation Framework — Hugging Face Space13 14This Gradio app evaluates and compares many AI agents across tasks (QA, summarization, reasoning...) using lightweight scorers and visualizations.15 16## How to use171. Upload a CSV/JSON/JSONL file with columns: `prompt`, `response`, `task`, `agent`, `reference` (reference optional).182. Click **Run Evaluation**.193. View per-task spider charts, heatmaps, bar plots in the Gallery, inspect per-example metrics in the table, and download the CSV report.20 21If no file is uploaded, a small synthetic demo dataset will be evaluated.22 23## Deploying24- Push this repo into a Hugging Face Space (Gradio). The `requirements.txt` will be installed automatically.25 26## Notes & Limitations27- Models used are lightweight but still require CPU memory (no Java).28- If `reference` is missing, hallucination/accuracy signals will be reduced.29- Coherence metric is a placeholder heuristic — you can replace it with grammar/perplexity models if desired.