README.md

---
library_name: transformers
license: apache-2.0
base_model: openai/whisper-tiny
language:
- orm
tags:
- automatic-speech-recognition
- african-languages
- waxal
- waxalnet
- orm
datasets:
- waxal-benchmarking/waxal
metrics:
- wer
- cer
---
# Whisper Tiny fine-tuned on WAXAL — Oromo

This model is part of **[WAXALNet](https://huggingface.co/waxal-benchmarking)**, a suite of ASR models fine-tuned on the [WAXAL corpus](https://huggingface.co/waxal-benchmarking) across 19 African languages, developed as part of the WAXAL ASR Benchmark study.

## Model Details

| | |
|---|---|
| **Language** | Oromo (`orm`) |
| **Language Family** | Afro-Asiatic |
| **Architecture** | Whisper Tiny (39M parameters) |
| **Base Model** | [openai/whisper-tiny](https://huggingface.co/openai/whisper-tiny) |
| **Training Data** | WAXAL corpus (conversational spontaneous speech) |
| **Test WER** | 29.3% |
| **Test CER** | 8.9% |
| **License** | apache-2.0 |

## Intended Use

This model is intended for automatic speech recognition of **Oromo** conversational speech. It was evaluated on the WAXAL test set (spontaneous, image-prompted speech) and partially on FLEURS (read speech). It is suitable for research and low-resource ASR applications. It is not recommended for high-stakes production use without further validation.

## Training Data

Fine-tuned on the [WAXAL corpus](https://huggingface.co/waxal-benchmarking), a large-scale dataset of transcribed, image-prompted spontaneous speech across 19 African languages recorded in participants' natural environments. The Oromo training split contains conversational speech across diverse speakers. Data is released under CC-BY 4.0.

## Usage

```python
from transformers import pipeline

asr = pipeline("automatic-speech-recognition",
               model="waxal-benchmarking/whisper-tiny-waxal-orm")
result = asr("audio.wav")
print(result["text"])
```

## Test Set Performance (WAXAL Benchmark)

Evaluated on the filtered WAXAL test set (duration >= 1.5s, speech rate >= 4 WPS).

| Metric | Score |
|---|---|
| **WER** | 29.3% |
| **CER** | 8.9% |

Full benchmark results across all 19 languages and 6 models are reported in the [WAXAL ASR Benchmark paper](https://arxiv.org/abs/2606.02375) (citation below).

## Training procedure

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 0.0001
- train_batch_size: 32
- eval_batch_size: 64
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 30
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch  | Step | Validation Loss | Wer    | Cer    |
|:-------------:|:------:|:----:|:---------------:|:------:|:------:|
| 0.6345        | 0.4195 | 500  | 0.6676          | 0.4803 | 0.1677 |
| 0.5221        | 0.8389 | 1000 | 0.5620          | 0.4046 | 0.1331 |
| 0.4200        | 1.2584 | 1500 | 0.5361          | 0.3759 | 0.1169 |
| 0.4088        | 1.6779 | 2000 | 0.5015          | 0.3635 | 0.1149 |
| 0.3198        | 2.0973 | 2500 | 0.4984          | 0.3467 | 0.1054 |
| 0.3386        | 2.5168 | 3000 | 0.4911          | 0.3410 | 0.1026 |
| 0.3245        | 2.9362 | 3500 | 0.4801          | 0.3401 | 0.1039 |
| 0.2734        | 3.3557 | 4000 | 0.4916          | 0.3408 | 0.1045 |
| 0.2750        | 3.7752 | 4500 | 0.4841          | 0.3374 | 0.1041 |
| 0.2215        | 4.1946 | 5000 | 0.5053          | 0.3413 | 0.1054 |


### Framework versions

- Transformers 5.0.0
- Pytorch 2.10.0+cu128
- Datasets 4.0.0
- Tokenizers 0.22.2

## Citation

```bibtex
@article{waxalnet2026,
  title  = {The WAXAL ASR Benchmark: Fine-Tuned Edge Models Across 19 African Languages},
  author = {Olufemi, Victor Tolulope and Babatunde, Oreoluwa and Njema, Ramsey and
             Gbotemi, Bolarinwa and Yen, Wanchi Lucia and Uzodinma, John and
             Ajayi, Sunday and Williams, Oluwademilade and Moshood, Kausar and
             Anyaele, Innocent Elendu and Arefaine, Akebert Tesfahunegn and
             Hunzwi, Candace and Daniel, Wongel Dawit and Namuganga, Emmilly Immaculate and
             Kadima, Cleophas and Bahizire, Athanase Biluge and Ranaivoson, Onitsiky and
             Aaron, Emmanuel and Ladislaus, Nicholaus Dismas and Muhammed, Idris and
             Simenya, Jonathan Enoch and Koome, Martin and Endaylalu, Matewos Tegete and
             Adeyemo, Peter Ifeoluwa and Birindwa, Hondi Prisca and Eze-Mbey, Ukachi Agnes and
             Oduro-Yeboah, Yacoba and Aremu, Toluwani and Adjovi, Pericles and
             Ngueajio, Mikel K and Mitra, Prasenjit},
  year   = {2026},
  note   = {arXiv preprint arXiv:2606.02375}
}
```

## Authors

Victor Tolulope Olufemi · Oreoluwa Babatunde · Ramsey Njema · Bolarinwa Gbotemi · Wanchi Lucia Yen · John Uzodinma · Sunday Ajayi · Oluwademilade Williams · Kausar Moshood · Innocent Elendu Anyaele · Akebert Tesfahunegn Arefaine · Candace Hunzwi · Wongel Dawit Daniel · Emmilly Immaculate Namuganga · Cleophas Kadima · Athanase Biluge Bahizire · Onitsiky Ranaivoson · Emmanuel Aaron · Nicholaus Dismas Ladislaus · Idris Muhammed · Jonathan Enoch Simenya · Martin Koome · Matewos Tegete Endaylalu · Peter Ifeoluwa Adeyemo · Hondi Prisca Birindwa · Ukachi Agnes Eze-Mbey · Yacoba Oduro-Yeboah · Toluwani Aremu · Pericles Adjovi · Mikel K Ngueajio · Prasenjit Mitra

## Acknowledgements

We thank the following contributors for their language expertise and native-speaker evaluation support:
Ajara Oyinloye, Abubakari Sadic Mohammed, Hafiz Adjei, Aliga Norah Lele, Marie-Louise B. Ndamuso, and Odong Diana.

This work was supported by **[Lynguallabs](https://lynguallabs.org/)** (compute, researchers & storage),
**[Open Token](https://opentoken.global/)** (compute resources), and
**[CMU Africa](https://www.africa.engineering.cmu.edu/)** (researchers & native speakers).