metadata
license: apache-2.0
base_model: facebook/wav2vec2-large-xlsr-53
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: wav2vec2-base-thai-5-google-colab
results: []
wav2vec2-base-thai-5-google-colab
This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.6198
- Wer: 0.5833
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: 16
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- 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: 500
- num_epochs: 20
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 5.8342 | 1.28 | 400 | 3.5303 | 1.0 |
| 3.0206 | 2.56 | 800 | 1.3367 | 0.9586 |
| 1.4665 | 3.84 | 1200 | 0.8291 | 0.8033 |
| 1.1346 | 5.12 | 1600 | 0.7229 | 0.7274 |
| 0.9699 | 6.4 | 2000 | 0.6384 | 0.7101 |
| 0.8754 | 7.68 | 2400 | 0.6393 | 0.6990 |
| 0.8001 | 8.96 | 2800 | 0.6427 | 0.6626 |
| 0.7116 | 10.24 | 3200 | 0.5992 | 0.6534 |
| 0.6743 | 11.52 | 3600 | 0.5936 | 0.6310 |
| 0.632 | 12.8 | 4000 | 0.6278 | 0.6348 |
| 0.5862 | 14.08 | 4400 | 0.6026 | 0.6152 |
| 0.563 | 15.36 | 4800 | 0.6020 | 0.5931 |
| 0.5359 | 16.64 | 5200 | 0.6278 | 0.5954 |
| 0.5065 | 17.92 | 5600 | 0.6273 | 0.5999 |
| 0.4958 | 19.2 | 6000 | 0.6198 | 0.5833 |
Framework versions
- Transformers 4.35.2
- Pytorch 2.1.0+cu121
- Datasets 2.16.1
- Tokenizers 0.15.0