metadata
license: apache-2.0
base_model: facebook/hubert-large-ls960-ft
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: hubert_large_531
results: []
hubert_large_531
This model is a fine-tuned version of facebook/hubert-large-ls960-ft on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.7843
- Wer: 0.2445
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: 0.0001
- train_batch_size: 2
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 8
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 1.1649 | 0.3378 | 100 | 0.9661 | 0.3229 |
| 1.2208 | 0.6757 | 200 | 0.8651 | 0.2918 |
| 1.412 | 1.0135 | 300 | 0.7681 | 0.2915 |
| 0.9419 | 1.3514 | 400 | 0.7712 | 0.2818 |
| 0.9161 | 1.6892 | 500 | 0.8276 | 0.2776 |
| 1.0943 | 2.0270 | 600 | 0.8074 | 0.2706 |
| 1.0036 | 2.3649 | 700 | 0.7917 | 0.2691 |
| 0.9262 | 2.7027 | 800 | 0.7897 | 0.2644 |
| 0.9841 | 3.0405 | 900 | 0.8006 | 0.2590 |
| 0.9327 | 3.3784 | 1000 | 0.8068 | 0.2630 |
| 0.7249 | 3.7162 | 1100 | 0.8546 | 0.2533 |
| 0.9704 | 4.0541 | 1200 | 0.8350 | 0.2533 |
| 0.8985 | 4.3919 | 1300 | 0.8414 | 0.2478 |
| 0.6858 | 4.7297 | 1400 | 0.8533 | 0.2469 |
| 1.0581 | 5.0676 | 1500 | 0.8200 | 0.2459 |
| 0.5899 | 5.4054 | 1600 | 0.8378 | 0.2459 |
| 0.6891 | 5.7432 | 1700 | 0.8094 | 0.2478 |
| 0.7276 | 6.0811 | 1800 | 0.8326 | 0.2438 |
| 0.5738 | 6.4189 | 1900 | 0.7844 | 0.2456 |
| 0.5837 | 6.7568 | 2000 | 0.7773 | 0.2473 |
| 0.5476 | 7.0946 | 2100 | 0.7725 | 0.2488 |
| 0.6004 | 7.4324 | 2200 | 0.7693 | 0.2471 |
| 0.5391 | 7.7703 | 2300 | 0.7843 | 0.2445 |
Framework versions
- Transformers 4.41.1
- Pytorch 2.3.0+cu121
- Datasets 2.19.1
- Tokenizers 0.19.1