metadata
license: apache-2.0
tags:
- generated_from_trainer
model-index:
- name: wav2vec_asr_swbd
results: []
wav2vec_asr_swbd
This model is a fine-tuned version of facebook/wav2vec2-large-robust-ft-swbd-300h on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.5214
- Wer: 0.7269
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: 0.0004
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 20
- total_train_batch_size: 80
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 150
- num_epochs: 20
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 3.0565 | 0.09 | 250 | 3.0007 | 1.0 |
| 1.4935 | 0.18 | 500 | 1.0356 | 0.8593 |
| 1.1937 | 0.26 | 750 | 0.7995 | 0.5973 |
| 1.0645 | 0.35 | 1000 | 0.6634 | 0.7798 |
| 0.9923 | 0.44 | 1250 | 0.6355 | 0.7755 |
| 0.9494 | 0.53 | 1500 | 0.6208 | 0.7647 |
| 0.9503 | 0.61 | 1750 | 0.5996 | 0.5880 |
| 0.9111 | 0.7 | 2000 | 0.5775 | 0.5689 |
| 0.9027 | 0.79 | 2250 | 0.5761 | 0.6729 |
| 0.8943 | 0.88 | 2500 | 0.5407 | 0.6768 |
| 0.8671 | 0.96 | 2750 | 0.5257 | 0.6955 |
| 0.823 | 1.05 | 3000 | 0.5520 | 0.8153 |
| 0.8279 | 1.14 | 3250 | 0.5214 | 0.7269 |
Framework versions
- Transformers 4.18.0
- Pytorch 1.11.0+cu113
- Datasets 1.18.4
- Tokenizers 0.11.6