metadata
license: apache-2.0
tags:
- generated_from_trainer
- robust-speech-event
datasets:
- common_voice
model-index:
- name: wav2vec2-large-xls-r-300m-as
results: []
wav2vec2-large-xls-r-300m-as
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: 0.8763
- Wer: 0.5679
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_ratio: 0.12
- num_epochs: 150
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 5.1465 | 25.0 | 400 | 3.3586 | 1.0 |
| 1.667 | 50.0 | 800 | 0.7541 | 0.6801 |
| 0.4061 | 75.0 | 1200 | 0.7686 | 0.6061 |
| 0.2274 | 100.0 | 1600 | 0.7842 | 0.5578 |
| 0.1479 | 125.0 | 2000 | 0.8560 | 0.5746 |
| 0.102 | 150.0 | 2400 | 0.8763 | 0.5679 |
Framework versions
- Transformers 4.16.2
- Pytorch 1.10.1+cu102
- Datasets 1.17.1.dev0
- Tokenizers 0.10.3