metadata
language:
- cy
license: apache-2.0
base_model: facebook/wav2vec2-large-xlsr-53
tags:
- automatic-speech-recognition
- python/custom_common_voice.py
- generated_from_trainer
datasets:
- custom_common_voice
metrics:
- wer
model-index:
- name: wav2vec2-xlsr-53-ft-ccv-en-cy
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: PYTHON/CUSTOM_COMMON_VOICE.PY - CY
type: custom_common_voice
config: cy
split: validation
args: 'Config: cy, Training split: train, Eval split: validation'
metrics:
- name: Wer
type: wer
value: 0.21777283505046477
wav2vec2-xlsr-53-ft-ccv-en-cy
This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on the PYTHON/CUSTOM_COMMON_VOICE.PY - CY dataset. It achieves the following results on the evaluation set:
- Loss: 0.2909
- Wer: 0.2178
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
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 5.8377 | 0.25 | 500 | 1.2190 | 0.8569 |
| 0.9829 | 0.51 | 1000 | 0.5585 | 0.4701 |
| 0.45 | 0.76 | 1500 | 0.4735 | 0.3901 |
| 0.3151 | 1.01 | 2000 | 0.4125 | 0.3418 |
| 0.2524 | 1.26 | 2500 | 0.3831 | 0.3117 |
| 0.243 | 1.52 | 3000 | 0.3661 | 0.3078 |
| 0.2341 | 1.77 | 3500 | 0.3489 | 0.2883 |
| 0.211 | 2.02 | 4000 | 0.3500 | 0.2738 |
| 0.1702 | 2.27 | 4500 | 0.3459 | 0.2704 |
| 0.1634 | 2.53 | 5000 | 0.3305 | 0.2583 |
| 0.1608 | 2.78 | 5500 | 0.3137 | 0.2479 |
| 0.1481 | 3.03 | 6000 | 0.3288 | 0.2562 |
| 0.1216 | 3.28 | 6500 | 0.3174 | 0.2446 |
| 0.1181 | 3.54 | 7000 | 0.3000 | 0.2325 |
| 0.1143 | 3.79 | 7500 | 0.2929 | 0.2326 |
| 0.1049 | 4.04 | 8000 | 0.2921 | 0.2218 |
| 0.0913 | 4.29 | 8500 | 0.2968 | 0.2208 |
| 0.0883 | 4.55 | 9000 | 0.2909 | 0.2178 |
Framework versions
- Transformers 4.33.3
- Pytorch 2.0.1+cu117
- Datasets 2.14.5
- Tokenizers 0.13.3