metadata
license: apache-2.0
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: my_awesome_asr_mind_model
results: []
my_awesome_asr_mind_model
This model is a fine-tuned version of facebook/wav2vec2-base on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.9568
- Wer: 1.0
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: 8
- 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_ratio: 0.1
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 6.3983 | 1.0 | 38 | 2.6746 | 1.0 |
| 1.0201 | 2.0 | 76 | 1.0183 | 1.0 |
| 0.98 | 3.0 | 114 | 1.2402 | 1.0 |
| 0.9409 | 4.0 | 152 | 0.9863 | 1.0 |
| 1.0142 | 5.0 | 190 | 0.9881 | 1.0 |
| 0.9615 | 6.0 | 228 | 0.9643 | 1.0 |
| 0.9693 | 7.0 | 266 | 0.9562 | 1.0 |
| 0.943 | 8.0 | 304 | 1.0394 | 1.0 |
| 0.9586 | 9.0 | 342 | 0.9711 | 1.0 |
| 0.9166 | 10.0 | 380 | 0.9568 | 1.0 |
Framework versions
- Transformers 4.24.0
- Pytorch 1.12.1+cu113
- Datasets 2.6.1
- Tokenizers 0.13.2