Model reference · open weights
romanian-wav2vec2 is an open-weight audio or speech model from gigant, 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
You can test this model online with the Space for Romanian Speech Recognition The model ranked TOP-1 on Romanian Speech Recognition during HuggingFace's Robust Speech Challenge : The 🤗 Speech Bench Speech Challenge Leaderboard Romanian Wav2Vec2 This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the Common Voice 8.0 - Romanian subset dataset, with extra training data from Romanian Speech Synthesis dataset. Without the 5-gram Language Model optimization, it achieves the following results on the evaluation set (Common Voice 8.0, Romanian subset, test split): - Loss: 0.1553 - Wer: 0.1174 - Cer: 0.0294 Model description The architecture is based on facebook/wav2vec2-xls-r-300m with a speech recognition CTC head and an added 5-gram language model (using pyctcdecode and kenlm) trained on the Romanian Corpora Parliament dataset. Those libraries are needed in order for the language model-boosted decoder to work. Intended uses & limitations The model is made for speech recognition in Romanian from audio clips sampled at 16kHz. The predicted text is lowercased and does not contain any punctuation. How to use Make sure you have installed the correct dependencies for the language model-boosted version to work. You can just run this command to install the kenlm and pyctcdecode libraries : from transformers import AutoProcessor, AutoModelForCTC processor = AutoProcessor.frompretrained("gigant/romanian-wav2vec2") model = AutoModelForCTC.frompretrained("gigant/romanian-wav2vec2") from transformers import pipeline asr = pipeline("automatic-speech-recognition", model="gigant/romanian-wav2vec2") from datasets import loaddataset dataset = loaddataset("gigant/romanianspeechsynthesis081") from IPython.display import Audio i = 0 sample = dataset["train"][i] Audio(sample["audio"]["array"], rate = sample["audio"]["samplingrate"]) import torchaudio import torch i = 0 audio = sample["audio"]["array"] rate = sample["audio"]["samplingrate"] resampler = torchaudio.transforms.Resample(rate, 16000) audio16 = resampler(torch.Tensor(audio)).numpy() Audio(audio16, rate=16000) predictedtext = asr(audio16) groundtruth = dataset["train"][i]["sentence"] print(f"Predicted text : {pre
Summarised from the published model card. Read the full card on the HuggingFace links below.
Specifications
| Maker | gigant |
|---|---|
| Type | Audio & music |
| Parameters (lead) | 315M |
| Variants | 1 |
| Runs with | transformers |
| Based on | facebook/wav2vec2-xls-r-300m |
| Released | 2022-03-02 |
| Popularity | 1.6M downloads / month |
| Likes | 8 |
| 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 |
|---|---|---|---|---|---|
| romanian-wav2vec2 | 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 | Robust Speech Event | Dev WER (without LM) | 46.99 |
| Automatic Speech Recognition | Robust Speech Event | Dev CER (without LM) | 16.04 |
| Automatic Speech Recognition | Robust Speech Event | Dev WER (with LM) | 38.63 |
| Automatic Speech Recognition | Robust Speech Event | Dev CER (with LM) | 14.52 |
| Automatic Speech Recognition | Common Voice | Test WER (without LM) | 11.73 |
| Automatic Speech Recognition | Common Voice | Test CER (without LM) | 2.93 |
| Automatic Speech Recognition | Common Voice | Test WER (with LM) | 7.31 |
| Automatic Speech Recognition | Common Voice | Test CER (with LM) | 2.17 |
| Automatic Speech Recognition | Robust Speech Event - Test Data | Test WER | 43.23 |
Using it via the API
Once AxForge deploys romanian-wav2vec2 for you, it answers on the OpenAI-compatible API — the same base URL and keys as every other model. (romanian-wav2vec2 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="romanian-wav2vec2" -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 ↗