distilhubert-finetuned-linnut-sm
This model is a fine-tuned version of ntu-spml/distilhubert on the codymd/linnut_audio_sm dataset. It achieves the following results on the evaluation set:
- Accuracy: 0.708
- F1: 0.5491
- Loss: 1.0891
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: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Accuracy | F1 | Validation Loss |
|---|---|---|---|---|---|
| 1.927 | 1.0 | 250 | 0.372 | 0.1179 | 2.1230 |
| 1.3331 | 2.0 | 500 | 0.568 | 0.2523 | 1.4067 |
| 1.3706 | 3.0 | 750 | 0.648 | 0.3386 | 1.1545 |
| 1.0341 | 4.0 | 1000 | 0.656 | 0.3842 | 1.1280 |
| 1.1953 | 5.0 | 1250 | 0.7 | 0.4663 | 1.0257 |
| 0.4731 | 6.0 | 1500 | 0.686 | 0.4635 | 1.0457 |
| 0.4189 | 7.0 | 1750 | 0.672 | 0.4710 | 1.0784 |
| 0.5092 | 8.0 | 2000 | 0.688 | 0.5073 | 1.0660 |
| 0.4302 | 9.0 | 2250 | 0.708 | 0.5301 | 1.0772 |
| 0.496 | 10.0 | 2500 | 0.708 | 0.5491 | 1.0891 |
Framework versions
- Transformers 4.48.3
- Pytorch 2.5.1+cu124
- Datasets 3.5.0
- Tokenizers 0.21.0