Model reference · open weights
wav2vec2-xls-r-hebrew is an open-weight audio or speech model from imvladikon, 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
should probably proofread and complete it, then remove this comment. -- wav2vec2-xls-r-300m-hebrew This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the private datasets in 2 stages - firstly was fine-tuned on a small dataset with good samples Then the obtained model was fine-tuned on a large dataset with the small good dataset, with various samples from different sources, and with an unlabeled dataset that was weakly labeled using a previously trained model. Small dataset: Large dataset: (weakly labeled data wasn't used in validation set) After firts training it achieves: on small dataset - Loss: 0.5438 - WER: 0.1773 on large dataset - WER: 0.3811 after second training: on small dataset - WER: 0.1697 on large dataset - Loss: 0.4502 - WER: 0.2318 Model description More information needed Intended uses & limitations More information needed Training and evaluation data More information needed Training procedure Training hyperparameters First training The following hyperparameters were used during training: - learningrate: 0.0003 - trainbatchsize: 8 - evalbatchsize: 8 - seed: 42 - distributedtype: multi-GPU - numdevices: 2 - gradientaccumulationsteps: 4 - totaltrainbatchsize: 64 - totalevalbatchsize: 16 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lrschedulertype: linear - lrschedulerwarmupsteps: 1000 - numepochs: 100.0 - mixedprecisiontraining: Native AMP Training results Second training The following hyperparameters were used during training: - learningrate: 0.0003 - trainbatchsize: 8 - evalbatchsize: 8 - seed: 42 - distributedtype: multi-GPU - numdevices: 2 - gradientaccumulationsteps: 4 - totaltrainbatchsize: 64 - totalevalbatchsize: 16 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lrschedulertype: linear - lrschedulerwarmupsteps: 1000 - numepochs: 60.0 - mixedprecisiontraining: Native AMP Training results Framework versions - Transformers 4.17.0.dev0 - Pytorch 1.10.2+cu102 - Datasets 1.18.2.dev0 - Tokenizers 0.11.0
Summarised from the published model card. Read the full card on the HuggingFace links below.
Specifications
| Maker | imvladikon |
|---|---|
| Type | Audio & music |
| Parameters (lead) | 315M |
| Variants | 1 |
| Runs with | transformers |
| Based on | facebook/wav2vec2-xls-r-300m |
| Released | 2022-03-02 |
| Popularity | 1.1M downloads / month |
| Likes | 6 |
| Licence | Commercial licence needed |
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-xls-r-300m-hebrew | 315M | BF16 | ~0.7 GB | ✓ | Weights ↗ |
Benchmarks
As published on the model card — the maker's own numbers, not measured by AxForge.
| Task | Dataset | Metric | Score |
|---|---|---|---|
| Automatic Speech Recognition | Custom Dataset | Test WER | 23.18 |
Using it via the API
Once AxForge deploys wav2vec2-xls-r-hebrew for you, it answers on the OpenAI-compatible API — the same base URL and keys as every other model. (wav2vec2-xls-r-hebrew 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="wav2vec2-xls-r-hebrew" -F file=@audio.mp3
Details
Languages
Tags
Licence
The weights are open but its licence needs a commercial agreement for business use. AxForge can arrange that licence and host the model for you — you pay AxForge, we settle with the model’s maker. Ask us for a quote. Read the licence ↗