|
--- |
|
license: apache-2.0 |
|
base_model: ntu-spml/distilhubert |
|
tags: |
|
- generated_from_trainer |
|
datasets: |
|
- marsyas/gtzan |
|
metrics: |
|
- accuracy |
|
- precision |
|
- recall |
|
- f1 |
|
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: 0.87 |
|
- name: Precision |
|
type: precision |
|
value: 0.8802816627816629 |
|
- name: Recall |
|
type: recall |
|
value: 0.87 |
|
- name: F1 |
|
type: f1 |
|
value: 0.8627110595989314 |
|
--- |
|
|
|
<!-- 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. --> |
|
|
|
[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="200" height="32"/>](https://wandb.ai/raspuntinov_ai/huggingface/runs/8epo656a) |
|
# 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: 0.6501 |
|
- Accuracy: 0.87 |
|
- Precision: 0.8803 |
|
- Recall: 0.87 |
|
- F1: 0.8627 |
|
|
|
## 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: 20 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 | |
|
|:-------------:|:-----:|:----:|:---------------:|:--------:|:---------:|:------:|:------:| |
|
| 2.1743 | 1.0 | 113 | 2.0604 | 0.38 | 0.5273 | 0.38 | 0.3101 | |
|
| 1.6179 | 2.0 | 226 | 1.4299 | 0.62 | 0.6136 | 0.62 | 0.5877 | |
|
| 1.0981 | 3.0 | 339 | 1.0223 | 0.79 | 0.8516 | 0.79 | 0.7669 | |
|
| 0.9785 | 4.0 | 452 | 0.8722 | 0.71 | 0.7748 | 0.71 | 0.6733 | |
|
| 0.8834 | 5.0 | 565 | 0.8363 | 0.76 | 0.7691 | 0.76 | 0.7449 | |
|
| 0.4936 | 6.0 | 678 | 0.6241 | 0.82 | 0.8313 | 0.82 | 0.8193 | |
|
| 0.2772 | 7.0 | 791 | 0.5648 | 0.85 | 0.8623 | 0.85 | 0.8459 | |
|
| 0.1213 | 8.0 | 904 | 0.6919 | 0.81 | 0.8429 | 0.81 | 0.7997 | |
|
| 0.0958 | 9.0 | 1017 | 0.5527 | 0.86 | 0.8682 | 0.86 | 0.8541 | |
|
| 0.0194 | 10.0 | 1130 | 0.6840 | 0.85 | 0.8645 | 0.85 | 0.8420 | |
|
| 0.0151 | 11.0 | 1243 | 0.6214 | 0.86 | 0.8642 | 0.86 | 0.8542 | |
|
| 0.1239 | 12.0 | 1356 | 0.6501 | 0.87 | 0.8803 | 0.87 | 0.8627 | |
|
| 0.0049 | 13.0 | 1469 | 0.6651 | 0.87 | 0.8803 | 0.87 | 0.8627 | |
|
| 0.0043 | 14.0 | 1582 | 0.7188 | 0.87 | 0.8803 | 0.87 | 0.8627 | |
|
| 0.0035 | 15.0 | 1695 | 0.6808 | 0.87 | 0.8803 | 0.87 | 0.8627 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.42.3 |
|
- Pytorch 2.1.2 |
|
- Datasets 2.20.0 |
|
- Tokenizers 0.19.1 |
|
|
