|
--- |
|
license: apache-2.0 |
|
base_model: facebook/wav2vec2-xls-r-2b |
|
tags: |
|
- automatic-speech-recognition |
|
- DewiBrynJones/banc-trawsgrifiadau-bangor-normalized |
|
- generated_from_trainer |
|
metrics: |
|
- wer |
|
model-index: |
|
- name: wav2vec2-xls-r-2b-ft-btb-cy |
|
results: [] |
|
--- |
|
|
|
<!-- This model card has been generated automatically according to the information the Trainer had access to. You |
|
should probably proofread and complete it, then remove this comment. --> |
|
|
|
# wav2vec2-xls-r-2b-ft-btb-cy |
|
|
|
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-2b](https://huggingface.co/facebook/wav2vec2-xls-r-2b) on the DEWIBRYNJONES/BANC-TRAWSGRIFIADAU-BANGOR-NORMALIZED - DEFAULT dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.3903 |
|
- Wer: 0.2957 |
|
|
|
## 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 |
|
- training_steps: 2500 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Wer | |
|
|:-------------:|:------:|:----:|:---------------:|:------:| |
|
| No log | 0.1414 | 100 | 1.2105 | 0.8709 | |
|
| No log | 0.2829 | 200 | 0.9787 | 0.6986 | |
|
| No log | 0.4243 | 300 | 1.1907 | 0.7127 | |
|
| No log | 0.5658 | 400 | 1.0559 | 0.7169 | |
|
| 1.4456 | 0.7072 | 500 | 1.2106 | 0.7944 | |
|
| 1.4456 | 0.8487 | 600 | 1.0232 | 0.7033 | |
|
| 1.4456 | 0.9901 | 700 | 1.0387 | 0.7336 | |
|
| 1.4456 | 1.1315 | 800 | 0.7234 | 0.5223 | |
|
| 1.4456 | 1.2730 | 900 | 0.7242 | 0.5566 | |
|
| 0.9155 | 1.4144 | 1000 | 0.7097 | 0.5259 | |
|
| 0.9155 | 1.5559 | 1100 | 0.6368 | 0.4797 | |
|
| 0.9155 | 1.6973 | 1200 | 0.6065 | 0.4653 | |
|
| 0.9155 | 1.8388 | 1300 | 0.6207 | 0.4717 | |
|
| 0.9155 | 1.9802 | 1400 | 0.5925 | 0.4707 | |
|
| 0.7436 | 2.1216 | 1500 | 0.5382 | 0.4046 | |
|
| 0.7436 | 2.2631 | 1600 | 0.5201 | 0.3996 | |
|
| 0.7436 | 2.4045 | 1700 | 0.4883 | 0.3698 | |
|
| 0.7436 | 2.5460 | 1800 | 0.4704 | 0.3659 | |
|
| 0.7436 | 2.6874 | 1900 | 0.4443 | 0.3521 | |
|
| 0.5645 | 2.8289 | 2000 | 0.4470 | 0.3476 | |
|
| 0.5645 | 2.9703 | 2100 | 0.4192 | 0.3242 | |
|
| 0.5645 | 3.1117 | 2200 | 0.4178 | 0.3161 | |
|
| 0.5645 | 3.2532 | 2300 | 0.4122 | 0.3054 | |
|
| 0.5645 | 3.3946 | 2400 | 0.3960 | 0.2990 | |
|
| 0.4232 | 3.5361 | 2500 | 0.3903 | 0.2957 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.40.2 |
|
- Pytorch 2.3.0+cu121 |
|
- Datasets 2.19.1 |
|
- Tokenizers 0.19.1 |
|
|
