metadata
language:
- ig
license: apache-2.0
base_model: openai/whisper-small
tags:
- generated_from_trainer
datasets:
- mozilla-foundation/common_voice_17_0
metrics:
- wer
model-index:
- name: Whisper Small Igbo
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: Common Voice 17
type: mozilla-foundation/common_voice_17_0
config: ig
split: test
args: ig
metrics:
- name: Wer
type: wer
value: 286.11111111111114
Whisper Small Igbo
This model is a fine-tuned version of openai/whisper-small on the Common Voice 17 dataset. It achieves the following results on the evaluation set:
- Loss: 8.1870
- Wer Ortho: 294.2857
- Wer: 286.1111
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: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: constant_with_warmup
- lr_scheduler_warmup_steps: 50
- training_steps: 4000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer Ortho | Wer |
|---|---|---|---|---|---|
| 0.0 | 500.0 | 500 | 3.2906 | 97.1429 | 88.8889 |
| 0.0 | 1000.0 | 1000 | 3.6536 | 91.4286 | 86.1111 |
| 0.0 | 1500.0 | 1500 | 4.3865 | 94.2857 | 91.6667 |
| 0.0 | 2000.0 | 2000 | 5.4348 | 102.8571 | 100.0 |
| 0.0 | 2500.0 | 2500 | 5.9169 | 94.2857 | 94.4444 |
| 0.0 | 3000.0 | 3000 | 6.6346 | 288.5714 | 277.7778 |
| 0.0 | 3500.0 | 3500 | 7.4267 | 297.1429 | 288.8889 |
| 0.0 | 4000.0 | 4000 | 8.1870 | 294.2857 | 286.1111 |
Framework versions
- Transformers 4.42.4
- Pytorch 2.3.1+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1