metadata
library_name: transformers
license: mit
base_model: facebook/w2v-bert-2.0
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: w2v-bert-2.0-CV_Fleurs-lg-5hrs-v5
results: []
w2v-bert-2.0-CV_Fleurs-lg-5hrs-v5
This model is a fine-tuned version of facebook/w2v-bert-2.0 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.4806
- Wer: 0.4843
- Cer: 0.1060
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: 0.0001
- train_batch_size: 4
- eval_batch_size: 2
- seed: 42
- optimizer: Use 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_ratio: 0.1
- num_epochs: 80
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
|---|---|---|---|---|---|
| 2.5225 | 1.0 | 515 | 0.4996 | 0.5506 | 0.1117 |
| 0.4435 | 2.0 | 1030 | 0.4619 | 0.4623 | 0.0969 |
| 0.3765 | 3.0 | 1545 | 0.4473 | 0.5063 | 0.1061 |
| 0.3573 | 4.0 | 2060 | 0.4596 | 0.4672 | 0.0962 |
| 0.3421 | 5.0 | 2575 | 0.4621 | 0.5073 | 0.1093 |
| 0.3235 | 6.0 | 3090 | 0.4548 | 0.5074 | 0.1057 |
| 0.3263 | 7.0 | 3605 | 0.4454 | 0.4664 | 0.1011 |
| 0.3125 | 8.0 | 4120 | 0.5261 | 0.5385 | 0.1251 |
| 0.2963 | 9.0 | 4635 | 0.4753 | 0.4890 | 0.1108 |
| 0.2527 | 10.0 | 5150 | 0.4803 | 0.4869 | 0.1085 |
| 0.2328 | 11.0 | 5665 | 0.4830 | 0.4710 | 0.1008 |
| 0.2077 | 12.0 | 6180 | 0.4806 | 0.4843 | 0.1060 |
Framework versions
- Transformers 4.46.1
- Pytorch 2.1.0+cu118
- Datasets 3.1.0
- Tokenizers 0.20.1