metadata
license: mit
Trained on:
- "train+validation" Swahili subsets of Common Voice 11.0
- "train+validation+test" Swahili subsets of Google Fleurs.
| GPU | Train Epochs | Training Loss | Validation Loss | WER | Train Time | Train Batch Size | Validation Batch Size |
|---|---|---|---|---|---|---|---|
| 1 x A100 | 1000 | 0.082000 | 19.887087 | 0.352631 | 11.2 hrs | 32 | 8 |