|
---
|
|
license: apache-2.0
|
|
base_model: facebook/hubert-base-ls960
|
|
tags:
|
|
- generated_from_trainer
|
|
metrics:
|
|
- accuracy
|
|
- precision
|
|
- recall
|
|
- f1
|
|
model-index:
|
|
- name: hubert-classifier
|
|
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
|
|
|
|
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: 1.1320
|
|
- Accuracy: 0.7724
|
|
- Precision: 0.8107
|
|
- Recall: 0.7724
|
|
- F1: 0.7633
|
|
- Binary: 0.8448
|
|
|
|
## 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: 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 | 0.17 | 50 | 4.2981 | 0.0242 | 0.0171 | 0.0242 | 0.0108 | 0.2760 |
|
|
| No log | 0.35 | 100 | 3.9571 | 0.0315 | 0.0038 | 0.0315 | 0.0059 | 0.3133 |
|
|
| No log | 0.52 | 150 | 3.7311 | 0.0678 | 0.0243 | 0.0678 | 0.0277 | 0.3424 |
|
|
| No log | 0.69 | 200 | 3.5663 | 0.0896 | 0.0557 | 0.0896 | 0.0513 | 0.3593 |
|
|
| No log | 0.86 | 250 | 3.4934 | 0.0944 | 0.0408 | 0.0944 | 0.0477 | 0.3545 |
|
|
| No log | 1.04 | 300 | 3.3705 | 0.1235 | 0.0864 | 0.1235 | 0.0706 | 0.3748 |
|
|
| No log | 1.21 | 350 | 3.2630 | 0.1429 | 0.1042 | 0.1429 | 0.0811 | 0.3906 |
|
|
| No log | 1.38 | 400 | 3.1551 | 0.1574 | 0.1413 | 0.1574 | 0.1118 | 0.4029 |
|
|
| No log | 1.55 | 450 | 3.0426 | 0.2349 | 0.1580 | 0.2349 | 0.1585 | 0.4593 |
|
|
| 3.6339 | 1.73 | 500 | 2.9462 | 0.2542 | 0.1854 | 0.2542 | 0.1806 | 0.4736 |
|
|
| 3.6339 | 1.9 | 550 | 2.8439 | 0.2663 | 0.2105 | 0.2663 | 0.2084 | 0.4814 |
|
|
| 3.6339 | 2.07 | 600 | 2.7192 | 0.3051 | 0.2797 | 0.3051 | 0.2411 | 0.5092 |
|
|
| 3.6339 | 2.24 | 650 | 2.6390 | 0.3293 | 0.3167 | 0.3293 | 0.2650 | 0.5274 |
|
|
| 3.6339 | 2.42 | 700 | 2.5491 | 0.3753 | 0.3731 | 0.3753 | 0.3290 | 0.5600 |
|
|
| 3.6339 | 2.59 | 750 | 2.4728 | 0.4092 | 0.3891 | 0.4092 | 0.3595 | 0.5823 |
|
|
| 3.6339 | 2.76 | 800 | 2.3395 | 0.4431 | 0.4205 | 0.4431 | 0.3917 | 0.6082 |
|
|
| 3.6339 | 2.93 | 850 | 2.2685 | 0.4552 | 0.4355 | 0.4552 | 0.4028 | 0.6160 |
|
|
| 3.6339 | 3.11 | 900 | 2.1883 | 0.4915 | 0.4680 | 0.4915 | 0.4387 | 0.6414 |
|
|
| 3.6339 | 3.28 | 950 | 2.1182 | 0.4843 | 0.5102 | 0.4843 | 0.4440 | 0.6363 |
|
|
| 2.6665 | 3.45 | 1000 | 2.0197 | 0.5448 | 0.5629 | 0.5448 | 0.5028 | 0.6804 |
|
|
| 2.6665 | 3.62 | 1050 | 1.9782 | 0.5327 | 0.5532 | 0.5327 | 0.4935 | 0.6712 |
|
|
| 2.6665 | 3.8 | 1100 | 1.9313 | 0.5593 | 0.5486 | 0.5593 | 0.5156 | 0.6930 |
|
|
| 2.6665 | 3.97 | 1150 | 1.8627 | 0.5908 | 0.5893 | 0.5908 | 0.5513 | 0.7119 |
|
|
| 2.6665 | 4.14 | 1200 | 1.8169 | 0.5908 | 0.5834 | 0.5908 | 0.5543 | 0.7128 |
|
|
| 2.6665 | 4.31 | 1250 | 1.7702 | 0.5835 | 0.5843 | 0.5835 | 0.5487 | 0.7077 |
|
|
| 2.6665 | 4.49 | 1300 | 1.7007 | 0.6344 | 0.6857 | 0.6344 | 0.6124 | 0.7438 |
|
|
| 2.6665 | 4.66 | 1350 | 1.6638 | 0.6199 | 0.6156 | 0.6199 | 0.5850 | 0.7354 |
|
|
| 2.6665 | 4.83 | 1400 | 1.6198 | 0.6368 | 0.6325 | 0.6368 | 0.6004 | 0.7482 |
|
|
| 2.6665 | 5.0 | 1450 | 1.5672 | 0.6804 | 0.6888 | 0.6804 | 0.6529 | 0.7753 |
|
|
| 2.0909 | 5.18 | 1500 | 1.5308 | 0.6683 | 0.6870 | 0.6683 | 0.6437 | 0.7692 |
|
|
| 2.0909 | 5.35 | 1550 | 1.4946 | 0.6877 | 0.6969 | 0.6877 | 0.6632 | 0.7811 |
|
|
| 2.0909 | 5.52 | 1600 | 1.4698 | 0.6755 | 0.6767 | 0.6755 | 0.6454 | 0.7743 |
|
|
| 2.0909 | 5.69 | 1650 | 1.4228 | 0.6804 | 0.7066 | 0.6804 | 0.6612 | 0.7785 |
|
|
| 2.0909 | 5.87 | 1700 | 1.3709 | 0.7312 | 0.7432 | 0.7312 | 0.7128 | 0.8140 |
|
|
| 2.0909 | 6.04 | 1750 | 1.3780 | 0.7215 | 0.7356 | 0.7215 | 0.7010 | 0.8082 |
|
|
| 2.0909 | 6.21 | 1800 | 1.3291 | 0.7215 | 0.7370 | 0.7215 | 0.7007 | 0.8090 |
|
|
| 2.0909 | 6.38 | 1850 | 1.3296 | 0.7191 | 0.7333 | 0.7191 | 0.7028 | 0.8056 |
|
|
| 2.0909 | 6.56 | 1900 | 1.3195 | 0.7191 | 0.7584 | 0.7191 | 0.7069 | 0.8048 |
|
|
| 2.0909 | 6.73 | 1950 | 1.2939 | 0.7191 | 0.7609 | 0.7191 | 0.7019 | 0.8065 |
|
|
| 1.75 | 6.9 | 2000 | 1.2800 | 0.7191 | 0.7353 | 0.7191 | 0.7018 | 0.8065 |
|
|
| 1.75 | 7.08 | 2050 | 1.2767 | 0.7094 | 0.7175 | 0.7094 | 0.6920 | 0.7998 |
|
|
| 1.75 | 7.25 | 2100 | 1.2280 | 0.7264 | 0.7689 | 0.7264 | 0.7148 | 0.8116 |
|
|
| 1.75 | 7.42 | 2150 | 1.2231 | 0.7385 | 0.7585 | 0.7385 | 0.7246 | 0.8201 |
|
|
| 1.75 | 7.59 | 2200 | 1.2198 | 0.7385 | 0.7563 | 0.7385 | 0.7248 | 0.8201 |
|
|
| 1.75 | 7.77 | 2250 | 1.1782 | 0.7482 | 0.7634 | 0.7482 | 0.7352 | 0.8269 |
|
|
| 1.75 | 7.94 | 2300 | 1.1848 | 0.7579 | 0.7900 | 0.7579 | 0.7519 | 0.8337 |
|
|
| 1.75 | 8.11 | 2350 | 1.1773 | 0.7579 | 0.7875 | 0.7579 | 0.7484 | 0.8346 |
|
|
| 1.75 | 8.28 | 2400 | 1.1752 | 0.7676 | 0.7965 | 0.7676 | 0.7594 | 0.8404 |
|
|
| 1.75 | 8.46 | 2450 | 1.1563 | 0.7724 | 0.8048 | 0.7724 | 0.7649 | 0.8438 |
|
|
| 1.5635 | 8.63 | 2500 | 1.1320 | 0.7724 | 0.8107 | 0.7724 | 0.7633 | 0.8448 |
|
|
| 1.5635 | 8.8 | 2550 | 1.1194 | 0.7700 | 0.8018 | 0.7700 | 0.7601 | 0.8421 |
|
|
| 1.5635 | 8.97 | 2600 | 1.1268 | 0.7554 | 0.7756 | 0.7554 | 0.7448 | 0.8329 |
|
|
| 1.5635 | 9.15 | 2650 | 1.1176 | 0.7676 | 0.7844 | 0.7676 | 0.7567 | 0.8404 |
|
|
|
|
|
|
### Framework versions
|
|
|
|
- Transformers 4.38.2
|
|
- Pytorch 2.3.0
|
|
- Datasets 2.19.1
|
|
- Tokenizers 0.15.1
|
|
|
