metadata
library_name: transformers
license: apache-2.0
base_model: facebook/wav2vec2-xls-r-300m
tags:
- generated_from_trainer
datasets:
- common_voice_17_0
metrics:
- wer
model-index:
- name: result_data-5
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: common_voice_17_0
type: common_voice_17_0
config: uk
split: test
args: uk
metrics:
- name: Wer
type: wer
value: 0.6674214548542315
result_data-5
This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the common_voice_17_0 dataset. It achieves the following results on the evaluation set:
- Loss: 0.4794
- Wer: 0.6674
- Cer: 0.2557
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: 8.442713223799316e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- distributed_type: multi-GPU
- num_devices: 2
- total_train_batch_size: 32
- total_eval_batch_size: 32
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 84
- num_epochs: 7.0
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
|---|---|---|---|---|---|
| 3.1606 | 0.9099 | 1000 | 3.1650 | 1.0 | 0.9920 |
| 1.2343 | 1.8198 | 2000 | 1.0356 | 0.9474 | 0.3937 |
| 0.7338 | 2.7298 | 3000 | 0.6668 | 0.7973 | 0.3038 |
| 0.6334 | 3.6397 | 4000 | 0.5813 | 0.7560 | 0.2852 |
| 0.5414 | 4.5496 | 5000 | 0.5283 | 0.6952 | 0.2675 |
| 0.5056 | 5.4595 | 6000 | 0.5042 | 0.6821 | 0.2633 |
| 0.4778 | 6.3694 | 7000 | 0.4794 | 0.6674 | 0.2557 |
Framework versions
- Transformers 4.49.0
- Pytorch 2.6.0+cu124
- Datasets 3.3.2
- Tokenizers 0.21.0