whisper-small-fo-fr API & Inference Endpoint

Credits (required)

Speech data: GbeBenin/FFSTC-2 (FFSTC 2: Extending the Fongbe to French Speech Translation Corpus).
Base model: OpenAI Whisper (openai/whisper-small, Apache-2.0).

Intended use

Table with columns: Direction, Supported
Direction	Supported
Fon audio → French text	Yes
Fon audio → Fon text	No
Other languages	No

Decode with French forced decoder / language=fr, task=transcribe (Whisper has no <|fon|> token; training used this setup).

Suggested generation kwargs (anti-hallucination):

python

dict(    language="fr",    task="transcribe",    num_beams=4,    length_penalty=0.8,    repetition_penalty=1.15,    no_repeat_ngram_size=3,    max_new_tokens=96,)

Training (summary)

Filtered FFSTC-2 French targets (drop address/parliament; upsample street).
Marian FR→EN→FR simplification when meaning preserved (sem_sim ≥ 80).
SpecAugment, cosine LR 2e-5, best checkpoint by MiniLM semantic similarity.
Early-stopped ~epoch 2 (best step 4400).

On-device pack

Sherpa-onnx int8 graphs: malinali-app/whisper-small-fo-fr-onnx.

Limitations

Literary / parliamentary FFSTC-2 style was down-weighted; street meaning preferred.
Gender/pronoun swaps and occasional topic drift still occur.
Single-speaker studio bias from the source corpus.

Citation

Please cite FFSTC-2 and Whisper if you use this model.

Credits (required)

Speech data: GbeBenin/FFSTC-2 (FFSTC 2: Extending the Fongbe to French Speech Translation Corpus).
Base model: OpenAI Whisper (openai/whisper-small, Apache-2.0).

Intended use

Table with columns: Direction, Supported
Direction	Supported
Fon audio → French text	Yes
Fon audio → Fon text	No
Other languages	No

Decode with French forced decoder / language=fr, task=transcribe (Whisper has no <|fon|> token; training used this setup).

Suggested generation kwargs (anti-hallucination):

python

dict(    language="fr",    task="transcribe",    num_beams=4,    length_penalty=0.8,    repetition_penalty=1.15,    no_repeat_ngram_size=3,    max_new_tokens=96,)

Training (summary)

Filtered FFSTC-2 French targets (drop address/parliament; upsample street).
Marian FR→EN→FR simplification when meaning preserved (sem_sim ≥ 80).
SpecAugment, cosine LR 2e-5, best checkpoint by MiniLM semantic similarity.
Early-stopped ~epoch 2 (best step 4400).

On-device pack

Sherpa-onnx int8 graphs: malinali-app/whisper-small-fo-fr-onnx.

Limitations

Literary / parliamentary FFSTC-2 style was down-weighted; street meaning preferred.
Gender/pronoun swaps and occasional topic drift still occur.
Single-speaker studio bias from the source corpus.

Citation

Please cite FFSTC-2 and Whisper if you use this model.

whisper-small-fo-fr

README

Credits (required)

Intended use

Training (summary)

On-device pack

Limitations

Citation

Explore FriendliAI today

README

Credits (required)

Intended use

Training (summary)

On-device pack

Limitations

Citation