---
license: apache-2.0
base_model: facebook/wav2vec2-large-xlsr-53
tags:
- automatic-speech-recognition
- DewiBrynJones/banc-trawsgrifiadau-bangor-normalized
- generated_from_trainer
metrics:
- wer
model-index:
- name: wav2vec2-xlsr-53-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-xlsr-53-ft-btb-cy

This model is a fine-tuned version of [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) on the DEWIBRYNJONES/BANC-TRAWSGRIFIADAU-BANGOR-NORMALIZED - DEFAULT dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3847
- Wer: 0.2967

## 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  | 3.7464          | 1.0    |
| No log        | 0.2829 | 200  | 2.9399          | 1.0    |
| No log        | 0.4243 | 300  | 2.5961          | 0.9991 |
| No log        | 0.5658 | 400  | 1.1619          | 0.7905 |
| 3.5448        | 0.7072 | 500  | 0.9466          | 0.6898 |
| 3.5448        | 0.8487 | 600  | 0.7895          | 0.6111 |
| 3.5448        | 0.9901 | 700  | 0.6820          | 0.5379 |
| 3.5448        | 1.1315 | 800  | 0.6039          | 0.4724 |
| 3.5448        | 1.2730 | 900  | 0.5631          | 0.4675 |
| 0.7808        | 1.4144 | 1000 | 0.5279          | 0.4291 |
| 0.7808        | 1.5559 | 1100 | 0.5024          | 0.3994 |
| 0.7808        | 1.6973 | 1200 | 0.4895          | 0.3895 |
| 0.7808        | 1.8388 | 1300 | 0.4596          | 0.3696 |
| 0.7808        | 1.9802 | 1400 | 0.4473          | 0.3611 |
| 0.6005        | 2.1216 | 1500 | 0.4332          | 0.3474 |
| 0.6005        | 2.2631 | 1600 | 0.4269          | 0.3418 |
| 0.6005        | 2.4045 | 1700 | 0.4155          | 0.3361 |
| 0.6005        | 2.5460 | 1800 | 0.4121          | 0.3214 |
| 0.6005        | 2.6874 | 1900 | 0.4145          | 0.3366 |
| 0.4666        | 2.8289 | 2000 | 0.3939          | 0.3114 |
| 0.4666        | 2.9703 | 2100 | 0.3889          | 0.3081 |
| 0.4666        | 3.1117 | 2200 | 0.3909          | 0.3064 |
| 0.4666        | 3.2532 | 2300 | 0.3874          | 0.3015 |
| 0.4666        | 3.3946 | 2400 | 0.3869          | 0.2983 |
| 0.3805        | 3.5361 | 2500 | 0.3847          | 0.2967 |


### Framework versions

- Transformers 4.40.2
- Pytorch 2.3.0+cu121
- Datasets 2.19.1
- Tokenizers 0.19.1