Downloads · 30 days
8
4% of all-time downloads
hatespeech_ast
DL-Project/hatespeech_ast
hatespeech_ast is a audio classification model from DL-Project. Use it for the audio classification task on the model card, and read the license before you ship it in a product. It is set up for transformers. The card lists the license as bsd-3-clause.
should probably proofread and complete it, then remove this comment. --
All-time downloads
213
Public
Parameters
86.2M
5.2 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors345 MB · 100%
At a glance
- Task
- Audio Classification
- Library
- transformers
- License
- bsd-3-clause
- Model type
- audio-spectrogram-transformer
- Access
- Public
- Created
- May 13, 2024
- Updated
- May 22, 2024
- SHA
- 59dec1f0
Base models
Model card
From the Hugging Face model README
<!-- 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. -->
hatespeech_ast
This model is a fine-tuned version of MIT/ast-finetuned-audioset-10-10-0.4593 on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.6306
- Accuracy: 0.6486
- Recall: 0.8368
- Precision: 0.6136
- F1: 0.7080
And the following results on the test set:
- Loss: 0.6441
- Accuracy: 0.6318
- Recall: 0.8191
- Precision: 0.6001
- F1: 0.6927
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: 16
- eval_batch_size: 16
- seed: 42
- gradient_accumulation_steps: 2
- 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 | Recall | Precision | F1 |
|---|---|---|---|---|---|---|---|
| 0.753 | 1.0 | 310 | 0.6793 | 0.5559 | 0.2258 | 0.6968 | 0.3411 |
| 0.6598 | 2.0 | 620 | 0.6447 | 0.6265 | 0.7575 | 0.6066 | 0.6737 |
| 0.6374 | 3.0 | 930 | 0.6306 | 0.6486 | 0.8368 | 0.6136 | 0.7080 |
| 0.5586 | 4.0 | 1240 | 0.7678 | 0.6091 | 0.9144 | 0.5727 | 0.7043 |
| 0.4008 | 5.0 | 1550 | 0.8134 | 0.6212 | 0.5515 | 0.6511 | 0.5972 |
| 0.2072 | 6.0 | 1860 | 1.0746 | 0.6265 | 0.7448 | 0.6088 | 0.6700 |
| 0.0904 | 7.0 | 2170 | 2.0297 | 0.6273 | 0.6878 | 0.6209 | 0.6526 |
| 0.0203 | 8.0 | 2480 | 3.0627 | 0.6236 | 0.6307 | 0.6302 | 0.6305 |
| 0.0244 | 9.0 | 2790 | 3.2017 | 0.6297 | 0.7013 | 0.6206 | 0.6585 |
| 0.0 | 10.0 | 3100 | 3.2659 | 0.6313 | 0.6331 | 0.6392 | 0.6361 |
Framework versions
- Transformers 4.40.2
- Pytorch 2.3.0+cu121
- Datasets 2.3.2
- Tokenizers 0.19.1
- Task
- Audio Classification
- Library
- transformers
- Type
- audio-spectrogram-transformer
- License
- bsd-3-clause
- Created
- May 13, 2024
- Updated
- May 22, 2024