metadata
license: apache-2.0
base_model: facebook/wav2vec2-large-xlsr-53
tags:
- automatic-speech-recognition
- techiaith/commonvoice_16_1_en_cy
- generated_from_trainer
metrics:
- wer
model-index:
- name: wav2vec2-xlsr-53-ft-ccv-en-cy
results: []
wav2vec2-xlsr-53-ft-ccv-en-cy
This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on the TECHIAITH/COMMONVOICE_16_1_EN_CY - DEFAULT dataset. It achieves the following results on the evaluation set:
- Loss: 0.2765
- Wer: 0.2115
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: 32
- eval_batch_size: 32
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 800
- training_steps: 9000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 5.9898 | 0.25 | 500 | 1.3093 | 0.7971 |
| 1.0749 | 0.5 | 1000 | 0.5816 | 0.4617 |
| 0.4332 | 0.75 | 1500 | 0.4834 | 0.4091 |
| 0.3303 | 1.01 | 2000 | 0.4203 | 0.3419 |
| 0.276 | 1.26 | 2500 | 0.3910 | 0.3186 |
| 0.2591 | 1.51 | 3000 | 0.3901 | 0.3067 |
| 0.2501 | 1.76 | 3500 | 0.3646 | 0.2895 |
| 0.224 | 2.01 | 4000 | 0.3517 | 0.2806 |
| 0.182 | 2.26 | 4500 | 0.3348 | 0.2656 |
| 0.1777 | 2.51 | 5000 | 0.3277 | 0.2612 |
| 0.1734 | 2.77 | 5500 | 0.3323 | 0.2643 |
| 0.1629 | 3.02 | 6000 | 0.3171 | 0.2485 |
| 0.1338 | 3.27 | 6500 | 0.3103 | 0.2398 |
| 0.1292 | 3.52 | 7000 | 0.2934 | 0.2268 |
| 0.1264 | 3.77 | 7500 | 0.2923 | 0.2248 |
| 0.118 | 4.02 | 8000 | 0.2880 | 0.2193 |
| 0.0996 | 4.27 | 8500 | 0.2793 | 0.2124 |
| 0.0969 | 4.52 | 9000 | 0.2765 | 0.2115 |
Framework versions
- Transformers 4.38.2
- Pytorch 2.2.1+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2