Spaces:
Running
Running
Commit History
-
feat: OpenEEGBench submission flow (arena runs the eval)
Replace the leftover LLM-leaderboard submission with a first-class OpenEEGBench flow. Users submit a model reference (model_cls + checkpoint + finetuning strategy/head/datasets); the backend validates it, enforces a 5-per-month per-user cap, and queues a request to braindecode/requests. A worker (stub) runs oeb.benchmark() and maps results to braindecode/contents. Backend: - submission_config: OEB enums/defaults (single source of truth) - submissions: pure validation, request building, monthly cap - results_mapping: OEB DataFrame -> contents schema (balanced_accuracy) - model_validation: EEG/OEB weights-source + license checks - OEB-shaped submit endpoint/service; eval worker stub (run_eval_worker.py) Frontend: - OEB submission form (model_cls, weights source, strategies, heads, PEFT modules, datasets, seeds, contact/visibility) with conditional fields - rewritten submission guide (5/month, help -> OpenEEGBench issues, citation) - queue/limit copy aligned; fixed duplicate style key…4db3be0
-
Fill model metadata in sample_results, drop unused fields
Fill #Params (M) with backbone-only param counts (head/final_layer excluded), instantiated from open-eeg-bench backbone configs: BIOT=3.19M, LaBraM=5.82M, CBraMod=4.92M, REVE=69.19M, EEGPT=null (no config in backbones.py). Set Hub License to bsd-3-clause for all entries. Drop embed_dim (dataset-dependent), trainable_params (dataset-dependent), and Weight type (Open LLM Leaderboard artefact, not relevant for EEG PEFT submissions) from the data and from every place that consumed them (transform_data, columnUtils, defaults.js, submission form, ModelService eval-queue path).8d3e1e1 -
Align internal dataset IDs with HF dataset IDs
Rename keys and accuracy_field values to match the HuggingFace dataset names under braindecode/, with dashes converted to underscores for Python/JS identifier safety: mat -> arithmetic_zyma2019 mumtaz -> mdd_mumtaz2016 seedv -> seed_v bcic-2020-3_accuracy-> bcic2020_3_accuracy996b915 -
Fix broken braindecode logo URL
The previous /_images/braindecode.svg path returns 404 on braindecode.org; the logo now lives under /_static/braindecode_long.svg.3d0b3bf -
Sync frontend dataset URL map with benchmark registry
30c59ab
-
Update main page title
Verified signature9e72865
-
Fix adapter method links not clickable by using MUI Link component
Replace Typography-based cell renderer with Link + Tooltip pattern (matching the architecture column) so PEFT doc links are clickable.cf8b1de -
Link adapter method column to HuggingFace PEFT documentation
Adapter methods (LoRA, AdaLoRA, DoRA, IA3, OFT) now link directly to their respective PEFT conceptual guide pages.90fca67 -
Fix dataset IDs to match actual HuggingFace repo names
Update bcic-2a → bcic2a, physionet-mi → physionet, chb-mit → chbmit in both the backend benchmark registry and frontend dataset URLs.af3dd25
-
Update frontend/public/index.html
Verified signature8d9d190
-
Add APA citations above BibTeX in About page
Each citation block now shows the APA-formatted reference first, followed by the BibTeX entry below. Both have independent copy buttons.d29d413 -
Remove Citations nav tab and route
The citation information is already available in the About page's Citation section, so the separate Citations page is redundant.b1aee4a -
Fix OOM build: load Plotly from CDN instead of npm bundle
The plotly.js-basic-dist npm package (~1.4MB gzipped) caused the Docker build to OOM on the HF Space cpu-basic tier. Replace it with a lightweight CDN loader that fetches plotly-basic at runtime. This removes the heavy webpack processing entirely and drops the JS bundle from 1.7MB to 250KB.ba02f4d -
Link dataset column headers to HF dataset pages
Each benchmark column header (BCIC-2a, PhysioNet, ISRUC, etc.) is now a hyperlink to its braindecode HF dataset page. The existing hover tooltip with dataset details is preserved. The link opens in a new tab and shows an underline on hover.31167c5 -
Link architecture column to braindecode documentation
Each foundation model name in the Architecture column now links to its braindecode API docs page (e.g. LaBraM → braindecode.models.Labram). Shows an external-link icon on hover.755439a -
Fix chart Y-axis: show architecture names and use automargin
Y-axis labels were clipped because model IDs are long and the left margin was fixed. Now shows the foundation model architecture name (since adapter type is already indicated by the grouping) with numeric suffixes for duplicates, and uses Plotly automargin so the axis sizes itself to fit the labels.3b7f8cf -
Add MOABB-style chart visualization to leaderboard
Add a Table/Chart toggle with a publication-grade dot plot view inspired by MOABB ranking figures. Models are grouped by adapter type with Tol's colorblind-friendly palette, hollow circle markers, dashed median reference line, and per-dataset selector buttons. Average view shows error bars (±1 std across benchmarks). Chart is lazy-loaded via React.lazy and the view mode persists in the URL (?view=chart).1155fb2 -
Add citation section to About page and widen Type column
Add OpenEEG-Bench and Braindecode BibTeX citations to the About page with copy-to-clipboard buttons. Widen the Type column from 65px to 120px to fit text labels like "Full Fine-tune".42cadaf -
Add braindecode logo side-by-side and remove adapter emojis
Display the braindecode logo next to the eegdash brain logo in the header, linking to braindecode.org. Replace all adapter method emoji icons with plain text labels across the leaderboard, about page, submission form, and filters.862c36c -
Fix SVG style attributes for React and add float32 to precision filter
- Convert style="fill:..." to style={{fill: "..."}} in HFLogo.js (React JSX) - Add float32 to FILTER_PRECISIONS default so EEG models are not filtered out - Update proxy to localhost:7861 for local developmentbf28da2 -
Fix server 500 error and replace HF logos with eegdash branding
- Add local sample data fallback when braindecode/contents HF dataset is unavailable (backend/app/data/sample_results.json with 20 entries) - Make fastapi-cache2 imports robust with graceful no-op fallback - Replace Hugging Face SVG logo with eegdash brain+EEG wave logo - Generate new favicon PNGs (32px, 256px) from eegdash SVG - Add missing http-proxy-middleware dependency to frontend package.json - Update OG URL to new Space name (OpenEEGBench) - Update AuthContainer to remove HF-specific branding text5b89071 -
Align frontend and backend with adapter_finetuning paper
Replace all LLM-specific references with real EEG benchmark data: - 8 EEG datasets (BCIC-2a, PhysioNet, ISRUC, TUAB, TUEV, CHB-MIT, FACED, SEED-V) - 7 adapter methods (LoRA, IA3, AdaLoRA, DoRA, OFT, Probe, Full Fine-tune) - 7 foundation models (LaBraM, EEGPT, BIOT, BENDR, SignalJEPA, CBraMod, REVE) - Remove official providers, MoE, flagged, merged features - Update params from billions to millions scale - Add real citations and dataset descriptions from the paper5763bf2 -
Migrate from Gradio to React + FastAPI architecture
Replace the old Gradio-based leaderboard with a React frontend and FastAPI backend, adapted from the Open LLM Leaderboard codebase for EEG model evaluation. Backend: FastAPI app with leaderboard, model submission, and voting endpoints. Frontend: React + MUI with leaderboard table, filters, model submission, voting, about page, and citations. Infrastructure: Dockerfile, docker-compose, and env configuration for HF Spaces.ee472e7