metadata
license: apache-2.0
tags:
- generated_from_trainer
base_model: facebook/wav2vec2-base
metrics:
- wer
model-index:
- name: krishivoice
results: []
krishivoice
This model is a fine-tuned version of facebook/wav2vec2-base on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 1.4807
- Wer: 0.7846
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: 1e-05
- train_batch_size: 20
- 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: 90
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 3.0074 | 11.3636 | 500 | 3.6612 | 1.0 |
| 2.9103 | 22.7273 | 1000 | 3.2488 | 1.0 |
| 2.8471 | 34.0909 | 1500 | 3.0912 | 1.0 |
| 2.2661 | 45.4545 | 2000 | 2.4130 | 0.9951 |
| 1.4007 | 56.8182 | 2500 | 1.9495 | 0.8854 |
| 1.0587 | 68.1818 | 3000 | 1.6888 | 0.8167 |
| 0.9269 | 79.5455 | 3500 | 1.4807 | 0.7846 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.3.0+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1