|
--- |
|
tags: |
|
- automatic-speech-recognition |
|
- timit_asr |
|
- generated_from_trainer |
|
datasets: |
|
- timit_asr |
|
model-index: |
|
- name: distilhubert-timit |
|
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. --> |
|
|
|
# distilhubert-timit |
|
|
|
This model is a fine-tuned version of [anton-l/distilhubert](https://huggingface.co/anton-l/distilhubert) on the TIMIT_ASR - NA dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 1.3688 |
|
- Wer: 0.6818 |
|
|
|
## 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: 1 |
|
- seed: 42 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- lr_scheduler_warmup_steps: 1000 |
|
- num_epochs: 20.0 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Wer | |
|
|:-------------:|:-----:|:----:|:---------------:|:------:| |
|
| 4.2247 | 0.69 | 100 | 3.8607 | 1.0 | |
|
| 2.9444 | 1.38 | 200 | 2.9509 | 1.0 | |
|
| 2.8858 | 2.07 | 300 | 2.8446 | 1.0 | |
|
| 2.2804 | 2.76 | 400 | 2.1985 | 1.0014 | |
|
| 1.505 | 3.45 | 500 | 1.4972 | 0.9609 | |
|
| 1.06 | 4.14 | 600 | 1.2014 | 0.8058 | |
|
| 1.0166 | 4.83 | 700 | 1.0605 | 0.7536 | |
|
| 0.966 | 5.52 | 800 | 0.9963 | 0.7101 | |
|
| 0.6857 | 6.21 | 900 | 0.9443 | 0.6898 | |
|
| 0.5859 | 6.9 | 1000 | 0.9043 | 0.6796 | |
|
| 0.6812 | 7.59 | 1100 | 0.9095 | 0.6716 | |
|
| 0.6088 | 8.28 | 1200 | 0.9422 | 0.6677 | |
|
| 0.4162 | 8.97 | 1300 | 0.9548 | 0.6657 | |
|
| 0.3411 | 9.66 | 1400 | 0.9901 | 0.6689 | |
|
| 0.3323 | 10.34 | 1500 | 0.9996 | 0.6638 | |
|
| 0.431 | 11.03 | 1600 | 1.0521 | 0.6708 | |
|
| 0.2029 | 11.72 | 1700 | 1.0946 | 0.6793 | |
|
| 0.1424 | 12.41 | 1800 | 1.1288 | 0.6712 | |
|
| 0.1922 | 13.1 | 1900 | 1.1456 | 0.6740 | |
|
| 0.326 | 13.79 | 2000 | 1.2077 | 0.6915 | |
|
| 0.0892 | 14.48 | 2100 | 1.2525 | 0.6796 | |
|
| 0.0769 | 15.17 | 2200 | 1.2313 | 0.6736 | |
|
| 0.0927 | 15.86 | 2300 | 1.3001 | 0.6864 | |
|
| 0.232 | 16.55 | 2400 | 1.3490 | 0.6963 | |
|
| 0.0485 | 17.24 | 2500 | 1.3268 | 0.6763 | |
|
| 0.0487 | 17.93 | 2600 | 1.3376 | 0.6780 | |
|
| 0.0607 | 18.62 | 2700 | 1.3701 | 0.6895 | |
|
| 0.1618 | 19.31 | 2800 | 1.3657 | 0.6796 | |
|
| 0.0415 | 20.0 | 2900 | 1.3688 | 0.6818 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.12.0.dev0 |
|
- Pytorch 1.8.1 |
|
- Datasets 1.14.1.dev0 |
|
- Tokenizers 0.10.3 |
|
|