|
--- |
|
license: apache-2.0 |
|
base_model: facebook/wav2vec2-base |
|
tags: |
|
- generated_from_trainer |
|
metrics: |
|
- wer |
|
model-index: |
|
- name: w2v2-base-pretrained_lr1e-4_at0.8_da0.3 |
|
results: [] |
|
--- |
|
|
|
<!-- This model card has been generated automatically according to the information the Trainer had access to. You |
|
should probably proofread and complete it, then remove this comment. --> |
|
|
|
# w2v2-base-pretrained_lr1e-4_at0.8_da0.3 |
|
|
|
This model is a fine-tuned version of [facebook/wav2vec2-base](https://huggingface.co/facebook/wav2vec2-base) on the None dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 1.6989 |
|
- Wer: 0.2119 |
|
|
|
## 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: 32 |
|
- eval_batch_size: 8 |
|
- seed: 42 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- lr_scheduler_warmup_steps: 100 |
|
- num_epochs: 200 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Wer | |
|
|:-------------:|:-----:|:----:|:---------------:|:------:| |
|
| 18.5074 | 7.14 | 100 | 3.4425 | 1.0 | |
|
| 3.1775 | 14.29 | 200 | 3.1511 | 1.0 | |
|
| 3.0847 | 21.43 | 300 | 3.1561 | 1.0 | |
|
| 2.9897 | 28.57 | 400 | 2.8890 | 1.0 | |
|
| 1.5122 | 35.71 | 500 | 1.0324 | 0.5404 | |
|
| 0.2161 | 42.86 | 600 | 1.2853 | 0.2960 | |
|
| 0.1074 | 50.0 | 700 | 1.4629 | 0.2456 | |
|
| 0.0707 | 57.14 | 800 | 1.5134 | 0.2183 | |
|
| 0.051 | 64.29 | 900 | 1.4349 | 0.2097 | |
|
| 0.04 | 71.43 | 1000 | 1.6989 | 0.2119 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.35.0 |
|
- Pytorch 2.0.0 |
|
- Datasets 2.14.6 |
|
- Tokenizers 0.14.1 |
|
|