cv26-hubert-svm-calibrated

CV26 HuBERT Calibrated SVM (Catalan Accent Oracle)

Research classifier artifact for Catalan Accent Oracle: five-way Catalan macro-dialect similarity scores from a pooled HuBERT embedding.

This is not geographic origin detection. Treat outputs as acoustic similarity to dialect areas.

Files

Labels (fixed order)

balearic, central, northern, northwestern, valencian

How to use with the app

# from the proj-accents repo root
mkdir -p models/cv26-hubert-svm-calibrated
hf download miquelt-9/cv26-hubert-svm-calibrated \
  --include "model.joblib" "metadata.json" \
  --local-dir models/cv26-hubert-svm-calibrated

Then start the FastAPI backend (uvicorn backend.app:app ...). On first inference, Transformers downloads the encoder BSC-LT/hubert-base-ca-2k into the local HF cache.

Pipeline

1. Encode audio with BSC-LT/hubert-base-ca-2k (frozen; not fine-tuned in this artifact)
2. Pool last hidden states with mean + std
3. Run this calibrated SVM → five dialect probabilities

Metrics (research snapshot)

Trained on 1,440 balanced Common Voice 26 Catalan clips (96 speakers × 3 clips × 5 dialects). Suitable for a local research prototype, not a polished public release.

License and attribution

• This repository / artifact (model.joblib + metadata.json): MIT. Sklearn head only; it does not redistribute HuBERT weights.
• Speech encoder BSC-LT/hubert-base-ca-2k: Apache License 2.0. Copyright (c) 2025 Language Technologies Unit, Barcelona Supercomputing Center. See the upstream model card and the project NOTICE.

If you later publish fine-tuned HuBERT checkpoints derived from BSC’s weights, those weights remain under Apache-2.0 and must keep BSC attribution.