Whisper Large Basque is an automatic speech recognition (ASR) model for Basque (eu) speech. It is fine-tuned from [openai/whisper-large] on the Basque portion of Mozilla Common Voice 13.0, achieving a Word Error Rate (WER) of 12.23% on the Common Voice evaluation split.

This model provides high-quality transcription for Basque speech, offering substantial improvements in accuracy over smaller Whisper variants while suitable for offline and batch processing tasks.

Model description

Architecture: Transformer-based encoder–decoder (Whisper)
Base model: openai/whisper-large
Language: Basque (eu)
Task: Automatic Speech Recognition (ASR)
Output: Text transcription in Basque
Decoding: Autoregressive sequence-to-sequence decoding

Leveraging Whisper’s multilingual pretraining, this large model is fine-tuned on Basque speech data to deliver highly accurate transcription for a low-resource language, suitable for research, media, and archival use cases.

Intended use

Primary use cases

High-quality transcription of Basque audio recordings
Offline or batch ASR pipelines
Research and development in Basque ASR
Media, educational, and archival transcription tasks

Intended users

Researchers working on Basque or low-resource ASR
Developers building Basque speech applications
Academic and institutional users

Out-of-scope use

Real-time or low-latency ASR without optimization
Speech translation tasks
Safety-critical applications without validation

Limitations and known issues

Performance may degrade on:
- Noisy or low-quality recordings
- Conversational or spontaneous speech
- Accents underrepresented in Common Voice
While highly accurate, transcription errors may still occur under challenging acoustic conditions
Dataset biases from Common Voice may be reflected in outputs

Users are encouraged to evaluate the model on their own data before deployment.

Training and evaluation data

Training data

Dataset: Mozilla Common Voice 13.0 (Basque subset)
Data type: Crowd-sourced, read speech
Preprocessing:
- Audio resampled to 16 kHz
- Text normalized using Whisper tokenizer
- Filtering of invalid or problematic samples

Evaluation data

Dataset: Mozilla Common Voice 13.0 (Basque evaluation split)
Metric: Word Error Rate (WER)

Evaluation results

Metric	Value
WER (eval)	12.23%

These results indicate state-of-the-art performance for Basque ASR using a large Whisper model.

Training procedure

Training hyperparameters

Learning rate: 1e-5
Optimizer: Adam (β1=0.9, β2=0.999, ε=1e-8)
LR scheduler: Linear
Warmup steps: 500
Training steps: 20,000
Train batch size: 32
Gradient accumulation steps: 2
Total effective batch size: 64
Evaluation batch size: 16
Seed: 42

Training results (summary)

Training Loss	Epoch	Step	Validation Loss	WER
0.0196	4.01	1000	0.2825	15.4725
0.0039	9.01	2000	0.3072	14.2270
0.0031	14.01	3000	0.3170	13.7652
0.0023	19.0	4000	0.3310	13.6640
0.0014	24.0	5000	0.3384	13.5749
0.0034	29.0	6000	0.3425	13.7450
0.0011	33.01	7000	0.3476	13.0990
0.001	38.01	8000	0.3432	13.0990
0.0004	43.01	9000	0.3524	12.8033
0.0017	48.01	10000	0.3620	13.3946
0.0003	53.0	11000	0.3564	12.6190
0.0001	58.0	12000	0.3675	12.6352
0.0	63.0	13000	0.3878	12.4286
0.0	67.01	14000	0.3996	12.3577
0.0	72.01	15000	0.4088	12.3456
0.0	77.01	16000	0.4167	12.3091
0.0	82.01	17000	0.4241	12.3112
0.0	87.0	18000	0.4302	12.3193
0.0	92.0	19000	0.4351	12.2565
0.0	97.0	20000	0.4369	12.2342

Framework versions

Transformers 4.33.0.dev0
PyTorch 2.0.1+cu117
Datasets 2.14.4
Tokenizers 0.13.3

How to use

python

from transformers import pipeline

hf_model = "HiTZ/whisper-large-eu"  # replace with actual repo ID
device = 0  # set to -1 for CPU

pipe = pipeline(
    task="automatic-speech-recognition",
    model=hf_model,
    device=device
)

result = pipe("audio.wav")
print(result["text"])

Ethical considerations and risks

This model transcribes speech and may process personal data.
Users should ensure compliance with applicable data protection laws (e.g., GDPR).
The model should not be used for surveillance or non-consensual audio processing.

Citation

If you use this model in your research, please cite:

bibtex

@misc{dezuazo2025whisperlmimprovingasrmodels,
  title={Whisper-LM: Improving ASR Models with Language Models for Low-Resource Languages},
  author={Xabier de Zuazo and Eva Navas and Ibon Saratxaga and Inma Hernáez Rioja},
  year={2025},
  eprint={2503.23542},
  archivePrefix={arXiv},
  primaryClass={cs.CL}
}

Please, check the related paper preprint in arXiv:2503.23542 for more details.

License

This model is available under the Apache-2.0 License. You are free to use, modify, and distribute this model as long as you credit the original creators.

Contact and attribution

Fine-tuning and evaluation: HiTZ/Aholab - Basque Center for Language Technology
Base model: OpenAI Whisper
Dataset: Mozilla Common Voice

For questions or issues, please open an issue in the model repository.

Funding

This project with reference 2022/TL22/00215335 has been parcially funded by the Ministerio de Transformación Digital and by the Plan de Recuperación, Transformación y Resiliencia – Funded by the European Union – NextGenerationEU ILENIA and by the project IkerGaitu funded by the Basque Government. This model was trained at Hyperion, one of the high-performance computing (HPC) systems hosted by the DIPC Supercomputing Center.