metadata
license: apache-2.0
base_model: facebook/wav2vec2-base
tags:
- generated_from_trainer
model-index:
- name: cmb-20s_asr-scr_w2v2-base_001
results: []
cmb-20s_asr-scr_w2v2-base_001
This model is a fine-tuned version of facebook/wav2vec2-base on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.4292
- Per: 0.1270
- Pcc: 0.6421
- Ctc Loss: 0.3927
- Mse Loss: 0.9759
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: 1e-05
- train_batch_size: 16
- eval_batch_size: 1
- seed: 1111
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 8928
- training_steps: 89280
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Per | Pcc | Ctc Loss | Mse Loss |
|---|---|---|---|---|---|---|---|
| 11.2433 | 3.0 | 8928 | 4.4453 | 0.9956 | 0.6128 | 3.7641 | 0.9015 |
| 3.0401 | 6.0 | 17856 | 1.4595 | 0.1725 | 0.6569 | 0.6144 | 0.8126 |
| 1.1033 | 9.0 | 26784 | 1.2737 | 0.1429 | 0.6630 | 0.4640 | 0.8414 |
| 0.6225 | 12.0 | 35712 | 1.2199 | 0.1361 | 0.6559 | 0.4317 | 0.9022 |
| 0.1917 | 15.0 | 44640 | 1.1453 | 0.1328 | 0.6507 | 0.4158 | 0.9433 |
| -0.2369 | 18.0 | 53568 | 1.0993 | 0.1299 | 0.6454 | 0.4055 | 1.0059 |
| -0.6422 | 21.0 | 62496 | 1.0154 | 0.1288 | 0.6420 | 0.4013 | 1.0500 |
| -1.0425 | 24.0 | 71424 | 0.7199 | 0.1279 | 0.6421 | 0.3942 | 1.0017 |
| -1.3918 | 27.0 | 80352 | 0.3882 | 0.1274 | 0.6428 | 0.3945 | 0.9264 |
| -1.6077 | 30.0 | 89280 | 0.4292 | 0.1270 | 0.6421 | 0.3927 | 0.9759 |
Framework versions
- Transformers 4.38.1
- Pytorch 2.0.1
- Datasets 2.16.1
- Tokenizers 0.15.2