metadata
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- common_voice_8_0
metrics:
- wer
model-index:
- name: wav2vec2-large-xls-r-300m-tr-colab
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: common_voice_8_0
type: common_voice_8_0
config: sw
split: test[:400]
args: sw
metrics:
- name: Wer
type: wer
value: 0.97
wav2vec2-large-xls-r-300m-tr-colab
This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the common_voice_8_0 dataset. It achieves the following results on the evaluation set:
- Loss: 0.4900
- Wer: 0.97
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: 50
- num_epochs: 5
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 6.5497 | 0.4 | 50 | 2.9819 | 1.0 |
| 2.8809 | 0.8 | 100 | 2.8873 | 1.0 |
| 2.8416 | 1.2 | 150 | 2.8427 | 1.0 |
| 2.8145 | 1.6 | 200 | 2.8067 | 1.0 |
| 2.747 | 2.0 | 250 | 2.7092 | 1.0 |
| 2.1095 | 2.4 | 300 | 1.3472 | 1.0 |
| 0.9546 | 2.8 | 350 | 0.7708 | 0.9975 |
| 0.6104 | 3.2 | 400 | 0.6317 | 0.9825 |
| 0.4941 | 3.6 | 450 | 0.5427 | 0.97 |
| 0.4345 | 4.0 | 500 | 0.5314 | 0.975 |
| 0.3327 | 4.4 | 550 | 0.4927 | 0.9625 |
| 0.3099 | 4.8 | 600 | 0.4900 | 0.97 |
Framework versions
- Transformers 4.27.4
- Pytorch 1.13.1+cu116
- Datasets 2.11.0
- Tokenizers 0.13.2