deepdml/whisper-medium-ig-mix-norm

Whisper Medium ig

This model is a fine-tuned version of openai/whisper-medium on the google/fleurs dataset. It achieves the following results on the evaluation set:

• Loss: 0.8789
• Wer: 39.4004
• Cer: 12.6049

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 1e-05
• train_batch_size: 32
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.04
• training_steps: 5000

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
0.107	0.2	1000	0.6957	43.6179	14.1782
0.0471	0.4	2000	0.7596	39.8086	12.9539
0.0273	0.6	3000	0.8375	40.2070	12.8450
0.0077	1.163	4000	0.8775	39.8814	12.8099
0.0149	1.363	5000	0.8789	39.4004	12.6049

Framework versions

• Transformers 4.42.0.dev0
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1

Citation

Please cite the model using the following BibTeX entry:

@misc{deepdml/whisper-medium-ig-mix-norm,
      title={Fine-tuned Whisper medium ASR model for speech recognition in Lingala},
      author={Jimenez, David},
      howpublished={\url{https://huggingface.co/deepdml/whisper-medium-ig-mix-norm}},
      year={2025}
    }