khyentsevision/tibetan-ai-leaderboard

Add a rebuildable lock file

Fix lint job: install git in python:3.12-slim before diffing

Require Google-style docstrings on public symbols via ruff

Add ruff lint gate scoped to changed files

Cut Translation Quality Metrics prose to a bare title line

Trim study methodology/findings prose pending peer review; add mitra-qwen35-2b-embedder metric

Translation Quality Metrics: strip LLM sub-tab prose, pills for judge rows

Add LLM-as-judge sub-tab: GEMBA-DA across twelve models

updated gitignore

Switch Tibetan Coverage to frequency-weighted scoring

Translation Quality Metrics: default sort by Pearson r, swap column order

Add 5 new reference-free metrics, Tibetan tokenizer coverage column

Rename MT Metrics tab to Translation Quality Metrics

MT Metrics: add ref-based/ref-free badges in All view, fix description

Add scripts/README.md: how to add metrics and models to MT Metrics tab

Add metric computation scripts from human-eval-data-study

Add All subtab to MT Metrics combining reference-based and reference-free

Refine MT Metrics tab: tooltips, remove category column, fix value display

Add MT Metrics tab and fix LanguageBench language pair description

Add GitHub link to DharmaBench tab

Add paper links to DharmaBench and TLUE tabs

Add LanguageBench description block and links to all tabs

Update header description and remove LanguageBench link

Add TNCC, WCM, MiLiC-Eval, TibetanQA tabs; add interpretive notes to all tabs

Add benchmark result files for TNCC, WCM, TibetanQA, TUSA/TNEC, MiLiC-Eval, Banzhida

Expand README with full benchmark descriptions and app architecture

Add Medal rankings and ScoreField bars to DharmaBench and TLUE tabs

Update About dialog to cover all three benchmark tabs

Move DharmaBench avg column left, expand benchmark descriptions

Add DharmaBench and TLUE benchmark tabs

Filter datasets and columns to Tibetan results only

Initial Tibetan-only leaderboard

initial commit