alibayram/mteb-turkish

26 models and 26 tasks done

27 models and 26 tasks done

before opus 4.5

embeddingmagibu200m added

Kaynakça eklendi

opus

Remove binary design files and update app with new features

Switch binary files to LFS

feat: Reimplement models view with advanced filtering and comparison, and restructure application layout with new navigation components.

windsurf first version

fix: update task reference paths for 'newmindai__TurkEmbed4STS' in models_manifest.json

feat: add new Turkish embedding model 'newmindai__TurkEmbed4STS' with multiple task references in models_manifest.json

refactor: update highlight styles and model identification for magibu in GeneralView and ModelsView

refactor: update model revision references in models_manifest.json for magibu embedding model

refactor: update model references in models_manifest.json to reflect migration from alibayram to magibu

refactor: migrate model files from alibayram to magibu, updating manifests and removing obsolete entries

fix: update model references and adjust related imports in the project

chore: rename model alibayram -> magibu and update manifests

fix: restore Download icon import in GeneralView

fix: exclude models missing scores for any selected task from the table

feat: add CSV download button for filtered table data

fix: translate all UI occurrences of 'Task' to 'Görev'

fix: restore missing getTaskType and getTaskDescription imports

feat: implement modal-based task selector to clean up sidebar

fix: re-add missing useEffect import in App.jsx

fix: adjust font sizes and correct model key for default task selection

fix: add missing imports (useMemo, icons) in GeneralView to resolve ReferenceErrors

fix: restore missing allTaskNames declaration and fix JSX syntax in GeneralView

fix: resolve SortIcon ReferenceError and improve UX compatibility

feat: refine task selection UI and logic

fix: resolve ReferenceError by adding missing TaskSelector import

feat: add task selector for user-customized model comparison

feat: implement truly fair scoring algorithm

build: deploy production build for static serving

fix: update YAML metadata for HuggingFace validation

refactor: clean up space structure

Add: Google EmbeddingGemma-300M results

Remove 'Sizin' badge from model names

Update: Source code with all General Table fixes

Fix: Working General Table with horizontal scroll, reordered columns, and no overlap

Fix: Use max-content width for proper horizontal scrolling

Fix: Remove fixed table layout and enforce column min-widths

Fix: Table layout, sticky column widths, and horizontal scrolling

Fix: Sanitize JSON NaN values and fix scrolling/import issues

Fix: Update source code backup with GeneralView fix

Fix: Import GeneralView and set as default view

Chore: Backup React source code

Feat: Add General Table view with highlighting for alibayram models

Refactor: Migrate to React + Vite for stability

Fix: Remove missing models from manifest to resolve 404s