<!-- 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. -->
hubert-large-ll60k-librispeech-clean-100h-demo-dist
This model is a fine-tuned version of facebook/hubert-base-ls960 on the LIBRISPEECH_ASR - CLEAN dataset. It achieves the following results on the evaluation set:
- Loss: 2.1361
- Wer: 0.9769
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
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 3.0
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 1.4699 | 0.15 | 100 | 2.0098 | 1.0 |
| 1.1 | 0.3 | 200 | 1.0910 | 0.9769 |
| 0.6468 | 0.45 | 300 | 1.5089 | 0.9769 |
| 0.6928 | 0.61 | 400 | 1.6171 | 1.0 |
| 0.698 | 0.76 | 500 | 1.2406 | 0.9769 |
| 1.0461 | 0.91 | 600 | 1.3760 | 1.0 |
| 0.6363 | 1.06 | 700 | 2.1654 | 1.0 |
| 0.6743 | 1.21 | 800 | 1.7481 | 0.9769 |
| 0.565 | 1.36 | 900 | 2.1965 | 1.0 |
| 0.5761 | 1.51 | 1000 | 1.8223 | 0.9769 |
| 0.6179 | 1.66 | 1100 | 1.9976 | 0.9769 |
| 0.5052 | 1.82 | 1200 | 1.5585 | 0.9769 |
| 0.5434 | 1.97 | 1300 | 2.0349 | 0.9769 |
| 0.4997 | 2.12 | 1400 | 2.4083 | 0.9769 |
| 0.489 | 2.27 | 1500 | 2.4164 | 0.9769 |
| 0.5104 | 2.42 | 1600 | 2.4970 | 0.9769 |
| 0.5324 | 2.57 | 1700 | 2.3352 | 0.9769 |
| 0.5207 | 2.72 | 1800 | 2.2009 | 0.9769 |
| 0.5224 | 2.87 | 1900 | 2.1035 | 0.9769 |
Framework versions
- Transformers 4.21.0.dev0
- Pytorch 1.12.0+cu113
- Datasets 2.3.2
- Tokenizers 0.12.1
