README.md

---
language:
- eu
license: apache-2.0
base_model: openai/whisper-base
tags:
- whisper-event
- generated_from_trainer
datasets:
- mozilla-foundation/common_voice_13_0
metrics:
- wer
model-index:
- name: Whisper Base Basque
  results:
  - task:
      name: Automatic Speech Recognition
      type: automatic-speech-recognition
    dataset:
      name: mozilla-foundation/common_voice_13_0 eu
      type: mozilla-foundation/common_voice_13_0
      config: eu
      split: test
      args: eu
    metrics:
    - name: Wer
      type: wer
      value: 25.977155818380655
---

# Whisper Base Basque

## Model summary

**Whisper Base Basque** is an automatic speech recognition (ASR) model for **Basque (eu)** speech. It is fine-tuned from [openai/whisper-base] on the **Basque portion of Mozilla Common Voice 13.0**, achieving a **Word Error Rate (WER) of 25.98%** on the Common Voice evaluation split.

The model offers a balance between model size and transcription quality, providing improved accuracy over tiny variants while remaining lighter than larger Whisper models.

---

## Model description

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

This model leverages Whisper’s multilingual pretraining and is fine-tuned on Basque speech data to enable ASR for a low-resource language with moderate computational requirements.

---

## Intended use

### Primary use cases

* Transcription of Basque audio recordings
* Offline or batch ASR pipelines
* Research on Basque and low-resource speech recognition
* Educational and archival transcription tasks

### Intended users

* Researchers working on Basque 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 additional validation

---

## Limitations and known issues

* Performance may degrade on:
  * Noisy or low-quality audio
  * Conversational or spontaneous speech
  * Accents underrepresented in Common Voice
* Accuracy is lower than larger Whisper variants
* Biases present in the Common Voice dataset 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) | **25.98%** |

These results reflect the trade-off between model compactness and transcription accuracy for Basque ASR.

---

## Training procedure

### Training hyperparameters

* Learning rate: 2.5e-5
* Optimizer: Adam (β1=0.9, β2=0.999, ε=1e-8)
* LR scheduler: Linear
* Warmup steps: 500
* Training steps: 5,000
* Train batch size: 128
* Evaluation batch size: 64
* Seed: 42

### Training results (summary)

| Training Loss | Epoch | Step | Validation Loss | WER     |
|:-------------:|:-----:|:----:|:---------------:|:-------:|
| 0.0174        | 9.01  | 1000 | 0.4597          | 27.3097 |
| 0.0016        | 19.01 | 2000 | 0.5160          | 26.0197 |
| 0.0007        | 29.0  | 3000 | 0.5520          | 25.9772 |
| 0.0005        | 38.02 | 4000 | 0.5728          | 26.1452 |
| 0.0004        | 48.01 | 5000 | 0.5818          | 26.2202 |

---

## 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-base-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](https://arxiv.org/abs/2503.23542)
for more details.

---

## License

This model is available under the
[Apache-2.0 License](https://www.apache.org/licenses/LICENSE-2.0).
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](https://proyectoilenia.es/) and by the project [IkerGaitu](https://www.hitz.eus/iker-gaitu/) funded by the Basque Government.
This model was trained at [Hyperion](https://scc.dipc.org/docs/systems/hyperion/overview/), one of the high-performance computing (HPC) systems hosted by the DIPC Supercomputing Center.