metadata
library_name: transformers
license: apache-2.0
base_model: facebook/wav2vec2-xls-r-300m
tags:
- generated_from_trainer
model-index:
- name: xls-r-300m-mocho-120
results: []
xls-r-300m-mocho-120
This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.0505
- Cer: 0.0300
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: 0.001
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- optimizer: Use OptimizerNames.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: 500
- num_epochs: 30
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Cer |
|---|---|---|---|---|
| 3.3995 | 3.6364 | 400 | 1.1680 | 0.4202 |
| 1.3772 | 7.2727 | 800 | 0.7910 | 0.3031 |
| 1.0304 | 10.9091 | 1200 | 0.5138 | 0.1994 |
| 0.8008 | 14.5455 | 1600 | 0.3396 | 0.1475 |
| 0.6073 | 18.1818 | 2000 | 0.2182 | 0.0974 |
| 0.4772 | 21.8182 | 2400 | 0.1480 | 0.0688 |
| 0.3533 | 25.4545 | 2800 | 0.0960 | 0.0500 |
| 0.2672 | 29.0909 | 3200 | 0.0505 | 0.0300 |
Framework versions
- Transformers 4.49.0.dev0
- Pytorch 2.5.1+cu124
- Datasets 3.2.0
- Tokenizers 0.21.0