Model reference · open weights
wav2vec2-large-xlsr-53-arabic is an open-weight audio or speech model from elgeish, listed in the AxForge catalogue. AxForge can bring it up on EU-owned hardware for you on request — with the licence handled where one is required.
About
Wav2Vec2-Large-XLSR-53-Arabic Fine-tuned facebook/wav2vec2-large-xlsr-53 on Arabic using the train splits of Common Voice and Arabic Speech Corpus. When using this model, make sure that your speech input is sampled at 16kHz. Usage The model can be used directly (without a language model) as follows: Here's the output: Evaluation The model can be evaluated as follows on the Arabic test data of Common Voice: Test Result: 26.55% Training For more details, see Fine-Tuning with Arabic Speech Corpus. This model represents Arabic in a format called Buckwalter transliteration. The Buckwalter format only includes ASCII characters, some of which are non-alpha (e.g., "" maps to "أ"). The lang-trans package is used to convert (transliterate) Arabic abjad. This script was used to first fine-tune facebook/wav2vec2-large-xlsr-53 on the train split of the Arabic Speech Corpus dataset; the test split was used for model selection; the resulting model at this point is saved as elgeish/wav2vec2-large-xlsr-53-levantine-arabic. Training was then resumed using the train split of the Common Voice dataset; the validation split was used for model selection; training was stopped to meet the deadline of Fine-Tune-XLSR Week: this model is the checkpoint at 100k steps and a validation WER of 23.39%. It's worth noting that validation WER is trending down, indicating the potential of further training (resuming the decaying learning rate at 7e-6). Future Work One area to explore is using attentionmask in model input, which is recommended here. Also, exploring data augmentation using datasets used to train models listed here.
Summarised from the published model card. Read the full card on the HuggingFace links below.
Specifications
| Maker | elgeish |
|---|---|
| Type | Audio & music |
| Variants | 1 |
| Runs with | transformers |
| Released | 2022-03-02 |
| Popularity | 732k downloads / month |
| Likes | 18 |
| Licence | Open weights |
How it works
Variants
Open weights ship in several sizes and precisions. One page, all the variants — pick the one that fits your GPU. VRAM figures are estimates from model size.
| Variant | Params | Precision | VRAM | Fits 16 GB | Weights |
|---|---|---|---|---|---|
| wav2vec2-large-xlsr-53-arabic | — | BF16 | — | — | Weights ↗ |
Benchmarks
As published on the model card — the maker's own numbers, not measured by AxForge.
| Task | Dataset | Metric | Score |
|---|---|---|---|
| Automatic Speech Recognition | Common Voice 6.1 (Arabic) | Test WER | 26.55 |
| Automatic Speech Recognition | Common Voice 6.1 (Arabic) | Validation WER | 23.39 |
Using it via the API
Once AxForge deploys elgeish-wav2vec2-large-xlsr-53-arabic for you, it answers on the OpenAI-compatible API — the same base URL and keys as every other model. (elgeish-wav2vec2-large-xlsr-53-arabic 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="elgeish-wav2vec2-large-xlsr-53-arabic" -F file=@audio.mp3
Details
Languages
Trained / evaluated on
Tags
Licence
Open weights under apache-2.0 — commercial use is permitted. Deploy it on AxForge EU hardware on request. Read the licence ↗