---
library_name: transformers
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: 0.84
---

<!-- 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. -->

# distilhubert-finetuned-gtzan

This model is a fine-tuned version of [ntu-spml/distilhubert](https://huggingface.co/ntu-spml/distilhubert) on the GTZAN dataset.
It achieves the following results on the evaluation set:
- Loss: 1.2032
- Accuracy: 0.84

## 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: 1
- eval_batch_size: 1
- 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 | Validation Loss | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| 1.3949        | 1.0   | 899  | 1.4476          | 0.61     |
| 1.1682        | 2.0   | 1798 | 1.4548          | 0.66     |
| 0.2096        | 3.0   | 2697 | 1.0660          | 0.72     |
| 3.073         | 4.0   | 3596 | 1.0223          | 0.78     |
| 0.0012        | 5.0   | 4495 | 0.9689          | 0.83     |
| 0.0088        | 6.0   | 5394 | 1.5253          | 0.77     |
| 0.0003        | 7.0   | 6293 | 0.9924          | 0.86     |
| 0.0002        | 8.0   | 7192 | 1.2847          | 0.85     |
| 0.0002        | 9.0   | 8091 | 1.2553          | 0.84     |
| 0.0002        | 10.0  | 8990 | 1.2032          | 0.84     |


### Framework versions

- Transformers 4.48.2
- Pytorch 2.5.1+cu121
- Datasets 3.2.0
- Tokenizers 0.21.0