metadata
license: apache-2.0
base_model: ntu-spml/distilhubert
tags:
- generated_from_trainer
datasets:
- birdclef/hubert
metrics:
- accuracy
model-index:
- name: distilhubert-finetuned-birdclef
results:
- task:
name: Audio Classification
type: audio-classification
dataset:
name: Birdclef 2024
type: birdclef/hubert
config: default
split: train
args: default
metrics:
- name: Accuracy
type: accuracy
value: 0.6974652493867539
distilhubert-finetuned-birdclef
This model is a fine-tuned version of ntu-spml/distilhubert on the Birdclef 2024 dataset. It achieves the following results on the evaluation set:
- Loss: 1.8592
- Accuracy: 0.6975
- F1 Macro: 0.4507
- F1 Weighted: 0.6871
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: 8
- eval_batch_size: 8
- seed: 42
- 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 Macro | F1 Weighted |
|---|---|---|---|---|---|---|
| 3.3212 | 1.0 | 2446 | 3.4889 | 0.2343 | 0.0453 | 0.1567 |
| 2.5841 | 2.0 | 4892 | 2.1989 | 0.5123 | 0.1673 | 0.4489 |
| 1.5152 | 3.0 | 7338 | 1.8349 | 0.5871 | 0.2449 | 0.5452 |
| 1.4771 | 4.0 | 9784 | 1.6815 | 0.6300 | 0.3213 | 0.6048 |
| 1.0287 | 5.0 | 12230 | 1.6218 | 0.6627 | 0.3498 | 0.6462 |
| 0.9425 | 6.0 | 14676 | 1.6177 | 0.6688 | 0.3835 | 0.6511 |
| 0.291 | 7.0 | 17122 | 1.7205 | 0.6832 | 0.3903 | 0.6682 |
| 0.244 | 8.0 | 19568 | 1.7817 | 0.6811 | 0.4049 | 0.6706 |
| 0.0593 | 9.0 | 22014 | 1.8653 | 0.6881 | 0.4282 | 0.6755 |
| 0.0754 | 10.0 | 24460 | 1.8877 | 0.6917 | 0.4320 | 0.6823 |
Framework versions
- Transformers 4.38.2
- Pytorch 2.2.1+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2