waxal-benchmarking/whisper-tiny-waxal-tir

Whisper Tiny fine-tuned on WAXAL — Tigrinya

This model is part of WAXALNet, a suite of ASR models fine-tuned on the WAXAL corpus across 19 African languages, developed as part of the WAXAL ASR Benchmark study.

Model Details

Language	Tigrinya (tir)
Language Family	Afro-Asiatic
Architecture	Whisper Tiny (39M parameters)
Base Model	openai/whisper-tiny
Training Data	WAXAL corpus (conversational spontaneous speech)
Test WER	60.3%
Test CER	43.2%
License	apache-2.0

Intended Use

This model is intended for automatic speech recognition of Tigrinya 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, a large-scale dataset of transcribed, image-prompted spontaneous speech across 19 African languages recorded in participants' natural environments. The Tigrinya training split contains conversational speech across diverse speakers. Data is released under CC-BY 4.0.

Usage

from transformers import pipeline

asr = pipeline("automatic-speech-recognition",
               model="waxal-benchmarking/whisper-tiny-waxal-tir")
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	60.3%
CER	43.2%

Full benchmark results across all 19 languages and 6 models are reported in the WAXAL ASR Benchmark paper (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
0.3859	0.4006	500	0.4361	0.8136	0.5189
0.2945	0.8013	1000	0.3107	0.7011	0.4381
0.2435	1.2019	1500	0.2828	0.6492	0.3937
0.2228	1.6026	2000	0.2577	0.6335	0.3917
0.2059	2.0032	2500	0.2464	0.6174	0.3888
0.2017	2.4038	3000	0.2356	0.5986	0.3700
0.1857	2.8045	3500	0.2304	0.5946	0.3726
0.1671	3.2051	4000	0.2286	0.5879	0.3688
0.1636	3.6058	4500	0.2207	0.5816	0.3611
0.1540	4.0064	5000	0.2225	0.5758	0.3624
0.1469	4.4071	5500	0.2176	0.5739	0.3584
0.1462	4.8077	6000	0.2171	0.5747	0.3622
0.1240	5.2083	6500	0.2239	0.5712	0.3601
0.1292	5.6090	7000	0.2217	0.5751	0.3624
0.1162	6.0096	7500	0.2243	0.5625	0.3546

Framework versions

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

Citation

@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   = {Preprint coming soon}
}

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 (compute, researchers & storage), Open Token (compute resources), and CMU Africa (researchers & native speakers).