|
--- |
|
license: apache-2.0 |
|
base_model: facebook/wav2vec2-base |
|
tags: |
|
- generated_from_trainer |
|
metrics: |
|
- wer |
|
model-index: |
|
- name: nyankole_wav2vec2 |
|
results: [] |
|
--- |
|
|
|
<!-- This model card has been generated automatically according to the information the Trainer had access to. You |
|
should probably proofread and complete it, then remove this comment. --> |
|
|
|
# nyankole_wav2vec2 |
|
|
|
This model is a fine-tuned version of [facebook/wav2vec2-base](https://huggingface.co/facebook/wav2vec2-base) on an unknown dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 2.8615 |
|
- Wer: 1.0 |
|
|
|
## 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: 0.0003 |
|
- train_batch_size: 4 |
|
- eval_batch_size: 8 |
|
- seed: 42 |
|
- gradient_accumulation_steps: 2 |
|
- total_train_batch_size: 8 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- lr_scheduler_warmup_steps: 150 |
|
- num_epochs: 30 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Wer | |
|
|:-------------:|:-------:|:----:|:---------------:|:---:| |
|
| 2.9097 | 0.9976 | 210 | 2.9031 | 1.0 | |
|
| 2.8649 | 2.0 | 421 | 2.8534 | 1.0 | |
|
| 2.8545 | 2.9976 | 631 | 2.8537 | 1.0 | |
|
| 2.8465 | 4.0 | 842 | 2.8554 | 1.0 | |
|
| 2.8509 | 4.9976 | 1052 | 2.8629 | 1.0 | |
|
| 2.8491 | 6.0 | 1263 | 2.8828 | 1.0 | |
|
| 2.8463 | 6.9976 | 1473 | 2.8570 | 1.0 | |
|
| 2.8477 | 8.0 | 1684 | 2.8675 | 1.0 | |
|
| 2.8478 | 8.9976 | 1894 | 2.8605 | 1.0 | |
|
| 2.8411 | 10.0 | 2105 | 2.8593 | 1.0 | |
|
| 2.8493 | 10.9976 | 2315 | 2.8573 | 1.0 | |
|
| 2.8478 | 12.0 | 2526 | 2.8564 | 1.0 | |
|
| 2.8823 | 12.9976 | 2736 | 2.8538 | 1.0 | |
|
| 2.8413 | 14.0 | 2947 | 2.8534 | 1.0 | |
|
| 2.8497 | 14.9976 | 3157 | 2.8487 | 1.0 | |
|
| 2.8439 | 16.0 | 3368 | 2.8642 | 1.0 | |
|
| 2.8442 | 16.9976 | 3578 | 2.8527 | 1.0 | |
|
| 2.8425 | 18.0 | 3789 | 2.8611 | 1.0 | |
|
| 2.841 | 18.9976 | 3999 | 2.8617 | 1.0 | |
|
| 2.8426 | 20.0 | 4210 | 2.8563 | 1.0 | |
|
| 2.8454 | 20.9976 | 4420 | 2.8527 | 1.0 | |
|
| 2.8396 | 22.0 | 4631 | 2.8568 | 1.0 | |
|
| 2.8449 | 22.9976 | 4841 | 2.8503 | 1.0 | |
|
| 2.8424 | 24.0 | 5052 | 2.8596 | 1.0 | |
|
| 2.8438 | 24.9976 | 5262 | 2.8624 | 1.0 | |
|
| 2.8414 | 26.0 | 5473 | 2.8606 | 1.0 | |
|
| 2.8387 | 26.9976 | 5683 | 2.8635 | 1.0 | |
|
| 2.8408 | 28.0 | 5894 | 2.8569 | 1.0 | |
|
| 2.8729 | 28.9976 | 6104 | 2.8640 | 1.0 | |
|
| 2.8417 | 29.9287 | 6300 | 2.8615 | 1.0 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.42.4 |
|
- Pytorch 2.1.0+cu118 |
|
- Datasets 2.20.0 |
|
- Tokenizers 0.19.1 |
|
|
