metadata
license: apache-2.0
base_model: ntu-spml/distilhubert
tags:
- generated_from_trainer
datasets:
- marsyas/gtzan
metrics:
- accuracy
model-index:
- name: distilhubert-finetuned-gtzan
results:
- task:
name: Audio Classification
type: audio-classification
dataset:
name: GTZAN
type: marsyas/gtzan
config: all
split: train
args: all
metrics:
- name: Accuracy
type: accuracy
value: 0.78
distilhubert-finetuned-gtzan
This model is a fine-tuned version of ntu-spml/distilhubert on the GTZAN dataset. It achieves the following results on the evaluation set:
- Loss: 0.7136
- Accuracy: 0.78
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: 5e-05
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 8
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 15
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 2.1794 | 1.0 | 112 | 2.1111 | 0.29 |
| 1.8527 | 2.0 | 225 | 1.7586 | 0.33 |
| 1.4409 | 3.0 | 337 | 1.4728 | 0.48 |
| 1.4484 | 4.0 | 450 | 1.3092 | 0.59 |
| 1.409 | 5.0 | 562 | 1.1073 | 0.68 |
| 1.0274 | 6.0 | 675 | 1.1107 | 0.67 |
| 1.0141 | 7.0 | 787 | 0.9898 | 0.7 |
| 0.8946 | 8.0 | 900 | 0.9540 | 0.68 |
| 1.0337 | 9.0 | 1012 | 0.8722 | 0.72 |
| 0.8065 | 10.0 | 1125 | 0.8196 | 0.74 |
| 1.0335 | 11.0 | 1237 | 0.7891 | 0.75 |
| 0.7712 | 12.0 | 1350 | 0.7298 | 0.78 |
| 0.7041 | 13.0 | 1462 | 0.7240 | 0.8 |
| 0.7287 | 14.0 | 1575 | 0.7206 | 0.78 |
| 0.9821 | 14.93 | 1680 | 0.7136 | 0.78 |
Framework versions
- Transformers 4.32.0.dev0
- Pytorch 2.0.1+cu117
- Datasets 2.13.1
- Tokenizers 0.13.3