|
--- |
|
license: apache-2.0 |
|
base_model: facebook/hubert-large-ll60k |
|
tags: |
|
- generated_from_trainer |
|
metrics: |
|
- wer |
|
model-index: |
|
- name: speech_ocean_hubert_mdd |
|
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. --> |
|
|
|
# speech_ocean_hubert_mdd |
|
|
|
This model is a fine-tuned version of [facebook/hubert-large-ll60k](https://huggingface.co/facebook/hubert-large-ll60k) on an unknown dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.2060 |
|
- Wer: 0.0531 |
|
- Cer: 0.0507 |
|
|
|
## 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: 16 |
|
- eval_batch_size: 8 |
|
- seed: 42 |
|
- gradient_accumulation_steps: 2 |
|
- total_train_batch_size: 32 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- lr_scheduler_warmup_steps: 500 |
|
- num_epochs: 20 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer | |
|
|:-------------:|:-------:|:----:|:---------------:|:------:|:------:| |
|
| 42.5933 | 0.9873 | 39 | 36.3096 | 0.9995 | 0.9985 | |
|
| 15.8463 | 2.0 | 79 | 7.9376 | 1.0 | 1.0 | |
|
| 6.5393 | 2.9873 | 118 | 4.5528 | 1.0 | 1.0 | |
|
| 3.9947 | 4.0 | 158 | 3.8592 | 1.0 | 1.0 | |
|
| 3.8025 | 4.9873 | 197 | 3.8002 | 1.0 | 1.0 | |
|
| 3.7706 | 6.0 | 237 | 3.7435 | 1.0 | 1.0 | |
|
| 3.7676 | 6.9873 | 276 | 3.7276 | 1.0 | 1.0 | |
|
| 3.7353 | 8.0 | 316 | 3.7150 | 1.0 | 1.0 | |
|
| 3.7126 | 8.9873 | 355 | 3.6717 | 1.0 | 1.0 | |
|
| 3.5628 | 10.0 | 395 | 3.3098 | 1.0 | 1.0 | |
|
| 2.837 | 10.9873 | 434 | 2.3304 | 0.8277 | 0.8915 | |
|
| 2.1018 | 12.0 | 474 | 1.5575 | 0.5441 | 0.6213 | |
|
| 1.6164 | 12.9873 | 513 | 1.0106 | 0.2678 | 0.2596 | |
|
| 1.1823 | 14.0 | 553 | 0.6938 | 0.1788 | 0.1509 | |
|
| 0.9451 | 14.9873 | 592 | 0.4673 | 0.1154 | 0.0925 | |
|
| 0.7055 | 16.0 | 632 | 0.3455 | 0.0893 | 0.0767 | |
|
| 0.5434 | 16.9873 | 671 | 0.2803 | 0.0718 | 0.0637 | |
|
| 0.4867 | 18.0 | 711 | 0.2362 | 0.0608 | 0.0566 | |
|
| 0.4172 | 18.9873 | 750 | 0.2125 | 0.0551 | 0.0522 | |
|
| 0.4406 | 19.7468 | 780 | 0.2060 | 0.0531 | 0.0507 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.40.0 |
|
- Pytorch 2.2.1+cu121 |
|
- Datasets 2.19.0 |
|
- Tokenizers 0.19.1 |
|
|
