Rework categories from a controlled vocabulary (via benchmarks table), drop removed per-model reported-category field, surface benchmark description/paper/impl links

Update for new dataset schema: flat benchmarks list + separate benchmarks metadata table

Adapt to simplified dataset schema (drop sources table, inline source link + benchmarks)

Adapt to simplified dataset schema (drop sources table, inline source link + benchmarks)

Switch benchmark usage chart from monthly to weekly buckets

Replace benchmark timeline with monthly usage bar chart

Document Latest releases tab (also refreshes dataset snapshot)

Add Latest releases landing tab: recent models with their benchmark suites and source links

Make dropdowns type-to-filter comboboxes (filter list, arrow/Enter/Escape keys, click to pick)

Rebuild UI with custom gr.HTML components (pills/select/slider/search, HTML tables, hand-rendered SVG charts); drop Plotly; fix duplicate-source model counting

Upload data.py with huggingface_hub

Upload app.py with huggingface_hub

Upload README.md with huggingface_hub

Upload requirements.txt with huggingface_hub

Upload data.py with huggingface_hub

Upload app.py with huggingface_hub

initial commit