metadata
license: apache-2.0
base_model: facebook/wav2vec2-xls-r-300m
tags:
- generated_from_trainer
datasets:
- common_voice
metrics:
- wer
model-index:
- name: Model_G_Wav2Vec2_Version3
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: common_voice
type: common_voice
config: id
split: test
args: id
metrics:
- name: Wer
type: wer
value: 0.3604325936975573
Model_G_Wav2Vec2_Version3
This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the common_voice dataset. It achieves the following results on the evaluation set:
- Loss: 0.4581
- Wer: 0.3604
- Cer: 0.0922
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: 30
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
|---|---|---|---|---|---|
| 3.8299 | 5.97 | 400 | 0.7799 | 0.6921 | 0.1952 |
| 0.3573 | 11.94 | 800 | 0.4949 | 0.4799 | 0.1235 |
| 0.1567 | 17.91 | 1200 | 0.4528 | 0.4228 | 0.1092 |
| 0.1011 | 23.88 | 1600 | 0.4695 | 0.3877 | 0.0993 |
| 0.0707 | 29.85 | 2000 | 0.4581 | 0.3604 | 0.0922 |
Framework versions
- Transformers 4.31.0
- Pytorch 2.0.1+cu117
- Datasets 1.18.3
- Tokenizers 0.13.3