|
--- |
|
license: apache-2.0 |
|
tags: |
|
- generated_from_trainer |
|
model-index: |
|
name: Waynehills-STT-doogie-server |
|
--- |
|
|
|
<!-- 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. --> |
|
|
|
# Waynehills-STT-doogie-server |
|
|
|
This model is a fine-tuned version of [Doogie/Waynehills-STT-doogie](https://huggingface.co/Doogie/Waynehills-STT-doogie) on an unknown dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 10.3564 |
|
- Wer: 1.0405 |
|
|
|
## 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: 8 |
|
- 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: 60 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Wer | |
|
|:-------------:|:-----:|:-----:|:---------------:|:------:| |
|
| 4.6722 | 1.92 | 1000 | 5.5301 | 1.0 | |
|
| 4.3024 | 3.84 | 2000 | 6.4368 | 1.0 | |
|
| 3.8135 | 5.76 | 3000 | 6.9063 | 1.0 | |
|
| 3.4163 | 7.68 | 4000 | 6.9737 | 1.0018 | |
|
| 3.1162 | 9.6 | 5000 | 7.1260 | 1.0027 | |
|
| 2.8724 | 11.52 | 6000 | 7.2143 | 1.0009 | |
|
| 2.6694 | 13.44 | 7000 | 7.4370 | 1.0050 | |
|
| 2.4808 | 15.36 | 8000 | 7.9850 | 1.0090 | |
|
| 2.2994 | 17.27 | 9000 | 8.1296 | 1.0198 | |
|
| 2.1436 | 19.19 | 10000 | 8.1327 | 1.0081 | |
|
| 2.0331 | 21.11 | 11000 | 8.2656 | 1.0135 | |
|
| 1.9278 | 23.03 | 12000 | 8.5640 | 1.0176 | |
|
| 1.8417 | 24.95 | 13000 | 8.5057 | 1.0212 | |
|
| 1.7496 | 26.87 | 14000 | 8.8110 | 1.0207 | |
|
| 1.6494 | 28.79 | 15000 | 9.0795 | 1.0306 | |
|
| 1.5882 | 30.71 | 16000 | 9.1341 | 1.0338 | |
|
| 1.5279 | 32.63 | 17000 | 9.2713 | 1.0284 | |
|
| 1.4712 | 34.55 | 18000 | 9.3591 | 1.0333 | |
|
| 1.4065 | 36.47 | 19000 | 9.4739 | 1.0293 | |
|
| 1.3637 | 38.39 | 20000 | 9.6498 | 1.0351 | |
|
| 1.3024 | 40.31 | 21000 | 9.7285 | 1.0365 | |
|
| 1.2737 | 42.23 | 22000 | 9.7353 | 1.0329 | |
|
| 1.2459 | 44.15 | 23000 | 10.0423 | 1.0374 | |
|
| 1.2079 | 46.07 | 24000 | 10.1164 | 1.0419 | |
|
| 1.1791 | 47.98 | 25000 | 10.1437 | 1.0437 | |
|
| 1.1593 | 49.9 | 26000 | 10.2292 | 1.0446 | |
|
| 1.1512 | 51.82 | 27000 | 10.2338 | 1.0405 | |
|
| 1.1041 | 53.74 | 28000 | 10.3070 | 1.0459 | |
|
| 1.1064 | 55.66 | 29000 | 10.3700 | 1.0419 | |
|
| 1.0783 | 57.58 | 30000 | 10.3950 | 1.0455 | |
|
| 1.0762 | 59.5 | 31000 | 10.3564 | 1.0405 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.12.5 |
|
- Pytorch 1.10.0+cu113 |
|
- Datasets 1.16.1 |
|
- Tokenizers 0.10.3 |
|
|
