|
---
|
|
license: apache-2.0
|
|
base_model: facebook/hubert-base-ls960
|
|
tags:
|
|
- generated_from_trainer
|
|
metrics:
|
|
- accuracy
|
|
- precision
|
|
- recall
|
|
- f1
|
|
model-index:
|
|
- name: hubert-classifier-aug
|
|
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. -->
|
|
|
|
# hubert-classifier-aug
|
|
|
|
This model is a fine-tuned version of [facebook/hubert-base-ls960](https://huggingface.co/facebook/hubert-base-ls960) on an unknown dataset.
|
|
It achieves the following results on the evaluation set:
|
|
- Loss: 0.5264
|
|
- Accuracy: 0.8625
|
|
- Precision: 0.8662
|
|
- Recall: 0.8625
|
|
- F1: 0.8517
|
|
- Binary: 0.9030
|
|
|
|
## 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: 32
|
|
- seed: 42
|
|
- gradient_accumulation_steps: 4
|
|
- total_train_batch_size: 128
|
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
|
- lr_scheduler_type: linear
|
|
- num_epochs: 10
|
|
- mixed_precision_training: Native AMP
|
|
|
|
### Training results
|
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 | Binary |
|
|
|:-------------:|:-----:|:----:|:---------------:|:--------:|:---------:|:------:|:------:|:------:|
|
|
| No log | 0.19 | 50 | 3.8945 | 0.0566 | 0.0077 | 0.0566 | 0.0127 | 0.3326 |
|
|
| No log | 0.38 | 100 | 3.4610 | 0.0701 | 0.0208 | 0.0701 | 0.0174 | 0.3418 |
|
|
| No log | 0.58 | 150 | 3.2223 | 0.1051 | 0.0294 | 0.1051 | 0.0364 | 0.3720 |
|
|
| No log | 0.77 | 200 | 3.1153 | 0.1294 | 0.0504 | 0.1294 | 0.0565 | 0.3795 |
|
|
| No log | 0.96 | 250 | 2.8292 | 0.1914 | 0.1010 | 0.1914 | 0.1073 | 0.4315 |
|
|
| No log | 1.15 | 300 | 2.7080 | 0.2264 | 0.1522 | 0.2264 | 0.1461 | 0.4496 |
|
|
| No log | 1.34 | 350 | 2.4083 | 0.2776 | 0.1986 | 0.2776 | 0.1896 | 0.4935 |
|
|
| No log | 1.53 | 400 | 2.2517 | 0.3720 | 0.2762 | 0.3720 | 0.2845 | 0.5580 |
|
|
| No log | 1.73 | 450 | 2.1201 | 0.3908 | 0.3501 | 0.3908 | 0.3098 | 0.5712 |
|
|
| 3.1927 | 1.92 | 500 | 1.9149 | 0.4582 | 0.3806 | 0.4582 | 0.3781 | 0.6210 |
|
|
| 3.1927 | 2.11 | 550 | 1.7920 | 0.5013 | 0.4684 | 0.5013 | 0.4456 | 0.6515 |
|
|
| 3.1927 | 2.3 | 600 | 1.5973 | 0.5418 | 0.4910 | 0.5418 | 0.4765 | 0.6803 |
|
|
| 3.1927 | 2.49 | 650 | 1.5067 | 0.5957 | 0.5572 | 0.5957 | 0.5409 | 0.7162 |
|
|
| 3.1927 | 2.68 | 700 | 1.3985 | 0.6253 | 0.6046 | 0.6253 | 0.5740 | 0.7361 |
|
|
| 3.1927 | 2.88 | 750 | 1.3198 | 0.6604 | 0.6224 | 0.6604 | 0.6114 | 0.7623 |
|
|
| 3.1927 | 3.07 | 800 | 1.2483 | 0.6685 | 0.6709 | 0.6685 | 0.6273 | 0.7674 |
|
|
| 3.1927 | 3.26 | 850 | 1.1560 | 0.7116 | 0.7063 | 0.7116 | 0.6710 | 0.7973 |
|
|
| 3.1927 | 3.45 | 900 | 1.0992 | 0.7197 | 0.7345 | 0.7197 | 0.6872 | 0.8030 |
|
|
| 3.1927 | 3.64 | 950 | 1.1148 | 0.7143 | 0.7477 | 0.7143 | 0.6918 | 0.7992 |
|
|
| 2.0117 | 3.84 | 1000 | 0.9688 | 0.7682 | 0.7634 | 0.7682 | 0.7404 | 0.8369 |
|
|
| 2.0117 | 4.03 | 1050 | 0.9990 | 0.7062 | 0.7148 | 0.7062 | 0.6717 | 0.7927 |
|
|
| 2.0117 | 4.22 | 1100 | 0.9516 | 0.7412 | 0.7619 | 0.7412 | 0.7229 | 0.8199 |
|
|
| 2.0117 | 4.41 | 1150 | 0.8740 | 0.7763 | 0.7947 | 0.7763 | 0.7582 | 0.8426 |
|
|
| 2.0117 | 4.6 | 1200 | 0.8611 | 0.7682 | 0.7800 | 0.7682 | 0.7469 | 0.8388 |
|
|
| 2.0117 | 4.79 | 1250 | 0.7992 | 0.7898 | 0.8228 | 0.7898 | 0.7775 | 0.8539 |
|
|
| 2.0117 | 4.99 | 1300 | 0.8161 | 0.7898 | 0.8209 | 0.7898 | 0.7756 | 0.8512 |
|
|
| 2.0117 | 5.18 | 1350 | 0.7420 | 0.7925 | 0.8144 | 0.7925 | 0.7768 | 0.8539 |
|
|
| 2.0117 | 5.37 | 1400 | 0.7420 | 0.7925 | 0.8070 | 0.7925 | 0.7712 | 0.8550 |
|
|
| 2.0117 | 5.56 | 1450 | 0.7126 | 0.8140 | 0.8187 | 0.8140 | 0.8017 | 0.8701 |
|
|
| 1.5617 | 5.75 | 1500 | 0.6797 | 0.8194 | 0.8436 | 0.8194 | 0.8086 | 0.8739 |
|
|
| 1.5617 | 5.94 | 1550 | 0.6877 | 0.8221 | 0.8279 | 0.8221 | 0.8028 | 0.8747 |
|
|
| 1.5617 | 6.14 | 1600 | 0.6547 | 0.8329 | 0.8525 | 0.8329 | 0.8230 | 0.8822 |
|
|
| 1.5617 | 6.33 | 1650 | 0.5935 | 0.8410 | 0.8589 | 0.8410 | 0.8270 | 0.8879 |
|
|
| 1.5617 | 6.52 | 1700 | 0.6423 | 0.8194 | 0.8255 | 0.8194 | 0.8052 | 0.8728 |
|
|
| 1.5617 | 6.71 | 1750 | 0.5980 | 0.8464 | 0.8610 | 0.8464 | 0.8322 | 0.8916 |
|
|
| 1.5617 | 6.9 | 1800 | 0.6111 | 0.8437 | 0.8543 | 0.8437 | 0.8287 | 0.8916 |
|
|
| 1.5617 | 7.09 | 1850 | 0.5835 | 0.8437 | 0.8588 | 0.8437 | 0.8336 | 0.8927 |
|
|
| 1.5617 | 7.29 | 1900 | 0.5804 | 0.8329 | 0.8461 | 0.8329 | 0.8210 | 0.8822 |
|
|
| 1.5617 | 7.48 | 1950 | 0.5711 | 0.8410 | 0.8580 | 0.8410 | 0.8290 | 0.8908 |
|
|
| 1.3255 | 7.67 | 2000 | 0.5468 | 0.8571 | 0.8633 | 0.8571 | 0.8457 | 0.9011 |
|
|
| 1.3255 | 7.86 | 2050 | 0.5384 | 0.8652 | 0.8720 | 0.8652 | 0.8553 | 0.9049 |
|
|
| 1.3255 | 8.05 | 2100 | 0.5673 | 0.8625 | 0.8684 | 0.8625 | 0.8547 | 0.9030 |
|
|
| 1.3255 | 8.25 | 2150 | 0.5450 | 0.8491 | 0.8582 | 0.8491 | 0.8403 | 0.8935 |
|
|
| 1.3255 | 8.44 | 2200 | 0.5278 | 0.8706 | 0.8770 | 0.8706 | 0.8630 | 0.9086 |
|
|
| 1.3255 | 8.63 | 2250 | 0.5339 | 0.8652 | 0.8692 | 0.8652 | 0.8542 | 0.9049 |
|
|
| 1.3255 | 8.82 | 2300 | 0.5469 | 0.8598 | 0.8648 | 0.8598 | 0.8489 | 0.9011 |
|
|
| 1.3255 | 9.01 | 2350 | 0.5404 | 0.8706 | 0.8747 | 0.8706 | 0.8602 | 0.9086 |
|
|
| 1.3255 | 9.2 | 2400 | 0.5455 | 0.8491 | 0.8565 | 0.8491 | 0.8378 | 0.8935 |
|
|
| 1.3255 | 9.4 | 2450 | 0.5317 | 0.8598 | 0.8664 | 0.8598 | 0.8479 | 0.9011 |
|
|
| 1.1934 | 9.59 | 2500 | 0.5227 | 0.8760 | 0.8798 | 0.8760 | 0.8657 | 0.9124 |
|
|
| 1.1934 | 9.78 | 2550 | 0.5278 | 0.8598 | 0.8653 | 0.8598 | 0.8481 | 0.9011 |
|
|
| 1.1934 | 9.97 | 2600 | 0.5264 | 0.8625 | 0.8662 | 0.8625 | 0.8517 | 0.9030 |
|
|
|
|
|
|
### Framework versions
|
|
|
|
- Transformers 4.38.2
|
|
- Pytorch 2.3.0
|
|
- Datasets 2.19.1
|
|
- Tokenizers 0.15.1
|
|
|
