metadata
language:
- tr
license: apache-2.0
tags:
- automatic-speech-recognition
- common_voice
- generated_from_trainer
datasets:
- common_voice
metrics:
- wer
model-index:
- name: wav2vec2-common_voice-tr-output
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: COMMON_VOICE - TR
type: common_voice
config: tr
split: test
args: 'Config: tr, Training split: train+validation, Eval split: test'
metrics:
- name: Wer
type: wer
value: 0.33888264732917983
wav2vec2-common_voice-tr-output
This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on the COMMON_VOICE - TR dataset. It achieves the following results on the evaluation set:
- Loss: 0.3805
- Wer: 0.3389
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
- num_epochs: 15.0
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Wer |
---|---|---|---|---|
No log | 0.92 | 100 | 3.6032 | 1.0 |
No log | 1.83 | 200 | 3.0158 | 0.9999 |
No log | 2.75 | 300 | 0.9692 | 0.8029 |
No log | 3.67 | 400 | 0.5820 | 0.6161 |
3.1812 | 4.59 | 500 | 0.4891 | 0.5095 |
3.1812 | 5.5 | 600 | 0.4719 | 0.4853 |
3.1812 | 6.42 | 700 | 0.4360 | 0.4539 |
3.1812 | 7.34 | 800 | 0.4098 | 0.4283 |
3.1812 | 8.26 | 900 | 0.4020 | 0.3993 |
0.2212 | 9.17 | 1000 | 0.4001 | 0.3806 |
0.2212 | 10.09 | 1100 | 0.4000 | 0.3873 |
0.2212 | 11.01 | 1200 | 0.4070 | 0.3751 |
0.2212 | 11.93 | 1300 | 0.3874 | 0.3551 |
0.2212 | 12.84 | 1400 | 0.3913 | 0.3561 |
0.0998 | 13.76 | 1500 | 0.3882 | 0.3492 |
0.0998 | 14.68 | 1600 | 0.3795 | 0.3403 |
Framework versions
- Transformers 4.28.1
- Pytorch 1.12.1+cu102
- Datasets 2.12.0
- Tokenizers 0.13.3