|
--- |
|
base_model: microsoft/unispeech-sat-base |
|
tags: |
|
- generated_from_trainer |
|
metrics: |
|
- accuracy |
|
- f1 |
|
- recall |
|
- precision |
|
model-index: |
|
- name: unispeech-sat-base-finetuned-common_voice |
|
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. --> |
|
|
|
# unispeech-sat-base-finetuned-common_voice |
|
|
|
This model is a fine-tuned version of [microsoft/unispeech-sat-base](https://huggingface.co/microsoft/unispeech-sat-base) on the None dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.0481 |
|
- Accuracy: 0.9925 |
|
- F1: 0.9925 |
|
- Recall: 0.9925 |
|
- Precision: 0.9928 |
|
- Mcc: 0.9907 |
|
- Auc: 0.9999 |
|
|
|
## 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: 3e-05 |
|
- train_batch_size: 8 |
|
- eval_batch_size: 8 |
|
- seed: 42 |
|
- gradient_accumulation_steps: 4 |
|
- total_train_batch_size: 32 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- lr_scheduler_warmup_ratio: 0.1 |
|
- num_epochs: 10 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Recall | Precision | Mcc | Auc | |
|
|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|:------:|:---------:|:------:|:------:| |
|
| 1.5302 | 1.0 | 50 | 1.4495 | 0.56 | 0.5047 | 0.5600 | 0.6655 | 0.4723 | 0.8635 | |
|
| 1.1592 | 2.0 | 100 | 0.9831 | 0.7125 | 0.6783 | 0.7125 | 0.7985 | 0.6723 | 0.9633 | |
|
| 0.7313 | 3.0 | 150 | 0.5535 | 0.9425 | 0.9428 | 0.9425 | 0.9455 | 0.9287 | 0.9926 | |
|
| 0.4431 | 4.0 | 200 | 0.2633 | 0.965 | 0.9651 | 0.9650 | 0.9676 | 0.9569 | 0.9976 | |
|
| 0.2353 | 5.0 | 250 | 0.1310 | 0.985 | 0.9850 | 0.985 | 0.9856 | 0.9814 | 0.9998 | |
|
| 0.1846 | 6.0 | 300 | 0.1136 | 0.9775 | 0.9775 | 0.9775 | 0.9783 | 0.9721 | 0.9978 | |
|
| 0.1464 | 7.0 | 350 | 0.0714 | 0.9875 | 0.9875 | 0.9875 | 0.9878 | 0.9844 | 1.0000 | |
|
| 0.1016 | 8.0 | 400 | 0.0592 | 0.99 | 0.9900 | 0.99 | 0.9902 | 0.9876 | 0.9999 | |
|
| 0.057 | 9.0 | 450 | 0.0466 | 0.9925 | 0.9925 | 0.9925 | 0.9928 | 0.9907 | 0.9999 | |
|
| 0.068 | 10.0 | 500 | 0.0481 | 0.9925 | 0.9925 | 0.9925 | 0.9928 | 0.9907 | 0.9999 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.40.2 |
|
- Pytorch 2.2.1+cu121 |
|
- Datasets 2.19.1 |
|
- Tokenizers 0.19.1 |
|
|
