metadata
license: apache-2.0
base_model: DewiBrynJones/wav2vec2-xlsr-53-ft-btb-ccv-cy
tags:
- automatic-speech-recognition
- DewiBrynJones/banc-trawsgrifiadau-bangor-clean
- generated_from_trainer
metrics:
- wer
model-index:
- name: wav2vec2-btb-ccv-ft-btb-cy
results: []
wav2vec2-btb-ccv-ft-btb-cy
This model is a fine-tuned version of DewiBrynJones/wav2vec2-xlsr-53-ft-btb-ccv-cy on the DEWIBRYNJONES/BANC-TRAWSGRIFIADAU-BANGOR-CLEAN - DEFAULT dataset. It achieves the following results on the evaluation set:
- Loss: 0.4376
- Wer: 0.3395
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: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 300
- training_steps: 3000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| No log | 0.1139 | 200 | 0.7528 | 0.5049 |
| No log | 0.2278 | 400 | 0.6966 | 0.5050 |
| 2.117 | 0.3417 | 600 | 0.6128 | 0.4761 |
| 2.117 | 0.4556 | 800 | 0.6332 | 0.5017 |
| 0.7606 | 0.5695 | 1000 | 0.5895 | 0.4577 |
| 0.7606 | 0.6834 | 1200 | 0.5553 | 0.4211 |
| 0.7606 | 0.7973 | 1400 | 0.5304 | 0.4196 |
| 0.7049 | 0.9112 | 1600 | 0.5061 | 0.3873 |
| 0.7049 | 1.0251 | 1800 | 0.5090 | 0.3959 |
| 0.6136 | 1.1390 | 2000 | 0.4839 | 0.3758 |
| 0.6136 | 1.2528 | 2200 | 0.4692 | 0.3659 |
| 0.6136 | 1.3667 | 2400 | 0.4569 | 0.3544 |
| 0.5388 | 1.4806 | 2600 | 0.4488 | 0.3485 |
| 0.5388 | 1.5945 | 2800 | 0.4411 | 0.3423 |
| 0.5275 | 1.7084 | 3000 | 0.4376 | 0.3395 |
Framework versions
- Transformers 4.44.0
- Pytorch 2.4.0+cu121
- Datasets 2.21.0
- Tokenizers 0.19.1