Model reference · open weights
wav2vec_asr_swbd_10_epochs is an open-weight audio or speech model from itaihay. AxForge deploys and operates it for you on dedicated EU-owned hardware — with the licence handled where one is required.
Available as managed deployment — configured and operated for you on dedicated EU hardware, quoted per deployment.
What it is
| Released by | itaihay |
|---|---|
| Type | Audio & music |
| Task | Speech→text |
| Runs with | transformers |
| Released | 2022-04-02 |
| Popularity | 733 downloads / month |
| Licence | Open weights |
About
This model is a fine-tuned version of facebook/wav2vec2-large-robust-ft-swbd-300h on an unknown dataset. It achieves the following results on the evaluation set:
More information needed
More information needed
More information needed
The following hyperparameters were used during training:
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 1.0682 | 0.22 | 5000 | 0.7383 | 0.4431 |
| 0.9143 | 0.44 | 10000 | 0.7182 | 0.4058 |
| 0.8905 | 0.66 | 15000 | 0.6291 | 0.3987 |
| 0.8354 | 0.87 | 20000 | 0.5976 | 0.3954 |
| 0.7749 | 1.09 | 25000 | 0.5773 | 0.3901 |
| 0.7336 | 1.31 | 30000 | 0.5812 | 0.3871 |
| 0.7314 | 1.53 | 35000 | 0.5802 | 0.3895 |
| 0.0 | 1.75 | 40000 | nan | 0.9627 |
| 0.0 | 1.97 | 45000 | nan | 0.9627 |
| 0.0 | 2.19 | 50000 | nan | 0.9627 |
| 0.0 | 2.4 | 55000 | nan | 0.9627 |
| 0.0 | 2.62 | 60000 | nan | 0.9627 |
| 0.0 | 2.84 | 65000 | nan | 0.9627 |
| 0.0 | 3.06 | 70000 | nan | 0.9627 |
| 0.0 | 3.28 | 75000 | nan | 0.9627 |
| 0.0 | 3.5 | 80000 | nan | 0.9627 |
| 0.0 | 3.72 | 85000 | nan | 0.9627 |
| 0.0 | 3.93 | 90000 | nan | 0.9627 |
| 0.0 | 4.15 | 95000 | nan | 0.9627 |
| 0.0 | 4.37 | 100000 | nan | 0.9627 |
| 0.0 | 4.59 | 105000 | nan | 0.9627 |
| 0.0 | 4.81 | 110000 | nan | 0.9627 |
| 0.0 | 5.03 | 115000 | nan | 0.9627 |
| 0.0 | 5.25 | 120000 | nan | 0.9627 |
| 0.0 | 5.46 | 125000 | nan | 0.9627 |
| 0.0 | 5.68 | 130000 | nan | 0.9627 |
| 0.0 | 5.9 | 135000 | nan | 0.9627 |
| 0.0 | 6.12 | 140000 | nan | 0.9627 |
| 0.0 | 6.34 | 145000 | nan | 0.9627 |
| 0.0 | 6.56 | 150000 | nan | 0.9627 |
| 0.0 | 6.78 | 155000 | nan | 0.9627 |
| 0.0 | 7.0 | 160000 | nan | 0.9627 |
| 0.0 | 7.21 | 165000 | nan | 0.9627 |
| 0.0 | 7.43 | 170000 | nan | 0.9627 |
| 0.0 | 7.65 | 175000 | nan | 0.9627 |
| 0.0 | 7.87 | 180000 | nan | 0.9627 |
| 0.0 | 8.09 | 185000 | nan | 0.9627 |
| 0.0 | 8.31 | 190000 | nan | 0.9627 |
| 0.0 | 8.53 | 195000 | nan | 0.9627 |
| 0.0 | 8.74 | 200000 | nan | 0.9627 |
| 0.0 | 8.96 | 205000 | nan | 0.9627 |
| 0.0 | 9.18 | 210000 | nan | 0.9627 |
| 0.0 | 9.4 | 215000 | nan | 0.9627 |
| 0.0 | 9.62 | 220000 | nan | 0.9627 |
| 0.0 | 9.84 | 225000 | nan | 0.9627 |
From the published model card. Full card on the HuggingFace links in the sidebar.
Using it via the API
Once AxForge deploys wav2vec-asr-swbd-10-epochs for you, it answers on the OpenAI-compatible API — the same base URL and keys as every other model. (wav2vec-asr-swbd-10-epochs below is illustrative; you get the exact model name on deployment.)
$ curl -sS https://api.axforge.ai/v1/audio/transcriptions \ -H "Authorization: Bearer $AXFORGE_API_KEY" \ -F model="wav2vec-asr-swbd-10-epochs" -F file=@audio.mp3
Create an account — your API key is available in the console. 3M free tokens every 30 days with every new account.