metadata
tags:
- generated_from_trainer
datasets:
- librispeech_asr
model-index:
- name: ''
results: []
This model was trained from scratch on the librispeech_asr dataset. It achieves the following results on the evaluation set:
- Loss: 12.7395
- Wer: 2.0272
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: 3e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- 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: 1000
- num_epochs: 20.0
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 6.6781 | 1.68 | 1500 | 6.6386 | 1.1672 |
| 6.6836 | 3.36 | 3000 | 6.6587 | 1.9518 |
| 6.6622 | 5.04 | 4500 | 6.5888 | 1.9276 |
| 5.844 | 6.73 | 6000 | 6.7220 | 1.9423 |
| 6.4588 | 8.41 | 7500 | 7.7569 | 1.9964 |
| 6.4097 | 10.09 | 9000 | 9.2515 | 2.0168 |
| 6.2676 | 11.77 | 10500 | 9.8159 | 2.0179 |
| 6.4948 | 13.45 | 12000 | 10.7091 | 2.0223 |
| 6.2728 | 15.13 | 13500 | 11.7747 | 2.0255 |
| 6.319 | 16.82 | 15000 | 12.2084 | 2.0259 |
| 5.8157 | 18.5 | 16500 | 12.7395 | 2.0272 |
Framework versions
- Transformers 4.17.0.dev0
- Pytorch 1.10.2+cu113
- Datasets 1.18.3
- Tokenizers 0.11.0