🚀 微調版DistilHuBERT模型用於GTZAN數據集
本項目的模型是基於ntu-spml/distilhubert在GTZAN數據集上進行微調得到的。該模型在評估集上取得了出色的效果，能夠為音頻相關任務提供有力支持。
🚀 快速開始
本模型是ntu-spml/distilhubert在GTZAN數據集上的微調版本。它在評估集上取得了以下結果：
🔧 技術細節
訓練超參數
訓練過程中使用了以下超參數：
- 學習率：4e-06
- 訓練批次大小：4
- 評估批次大小：4
- 隨機種子：42
- 梯度累積步數：2
- 總訓練批次大小：8
- 優化器：Adam（β1=0.9，β2=0.999，ε=1e-08）
- 學習率調度器類型：線性
- 學習率調度器預熱比例：0.1
- 訓練輪數：50
訓練結果
| 訓練損失 |
輪數 |
步數 |
驗證損失 |
準確率 |
| 0.2359 |
1.0 |
112 |
0.4776 |
0.87 |
| 0.1235 |
2.0 |
225 |
0.4872 |
0.84 |
| 0.2083 |
3.0 |
337 |
0.4910 |
0.85 |
| 0.19 |
4.0 |
450 |
0.4953 |
0.87 |
| 0.1128 |
5.0 |
562 |
0.4801 |
0.87 |
| 0.1644 |
6.0 |
675 |
0.4703 |
0.87 |
| 0.0699 |
7.0 |
787 |
0.4692 |
0.85 |
| 0.1082 |
8.0 |
900 |
0.4708 |
0.87 |
| 0.0898 |
9.0 |
1012 |
0.4347 |
0.89 |
| 0.1071 |
10.0 |
1125 |
0.5310 |
0.85 |
| 0.0727 |
11.0 |
1237 |
0.4765 |
0.87 |
| 0.0338 |
12.0 |
1350 |
0.4859 |
0.87 |
| 0.0233 |
13.0 |
1462 |
0.4713 |
0.87 |
| 0.0248 |
14.0 |
1575 |
0.5068 |
0.88 |
| 0.0263 |
15.0 |
1687 |
0.4874 |
0.88 |
| 0.0185 |
16.0 |
1800 |
0.4925 |
0.88 |
| 0.0142 |
17.0 |
1912 |
0.4766 |
0.89 |
| 0.0178 |
18.0 |
2025 |
0.4850 |
0.89 |
| 0.0153 |
19.0 |
2137 |
0.4660 |
0.88 |
| 0.012 |
20.0 |
2250 |
0.4831 |
0.88 |
| 0.0113 |
21.0 |
2362 |
0.4965 |
0.89 |
| 0.0106 |
22.0 |
2475 |
0.5098 |
0.89 |
| 0.011 |
23.0 |
2587 |
0.5093 |
0.89 |
| 0.009 |
24.0 |
2700 |
0.4989 |
0.91 |
| 0.0094 |
25.0 |
2812 |
0.4999 |
0.89 |
| 0.0441 |
26.0 |
2925 |
0.5197 |
0.88 |
| 0.0079 |
27.0 |
3037 |
0.5115 |
0.89 |
| 0.0072 |
28.0 |
3150 |
0.5136 |
0.88 |
| 0.007 |
29.0 |
3262 |
0.5394 |
0.88 |
| 0.0068 |
30.0 |
3375 |
0.5374 |
0.88 |
| 0.0061 |
31.0 |
3487 |
0.5221 |
0.88 |
| 0.0533 |
32.0 |
3600 |
0.5775 |
0.87 |
| 0.0055 |
33.0 |
3712 |
0.5632 |
0.88 |
| 0.0059 |
34.0 |
3825 |
0.5584 |
0.87 |
| 0.0051 |
35.0 |
3937 |
0.5444 |
0.88 |
| 0.0051 |
36.0 |
4050 |
0.5373 |
0.88 |
| 0.0045 |
37.0 |
4162 |
0.5723 |
0.87 |
| 0.0058 |
38.0 |
4275 |
0.5773 |
0.87 |
| 0.0043 |
39.0 |
4387 |
0.5455 |
0.88 |
| 0.0044 |
40.0 |
4500 |
0.5686 |
0.88 |
| 0.004 |
41.0 |
4612 |
0.5622 |
0.87 |
| 0.004 |
42.0 |
4725 |
0.5797 |
0.88 |
| 0.0042 |
43.0 |
4837 |
0.5621 |
0.88 |
| 0.0037 |
44.0 |
4950 |
0.5734 |
0.87 |
| 0.0048 |
45.0 |
5062 |
0.5774 |
0.88 |
| 0.0039 |
46.0 |
5175 |
0.5901 |
0.87 |
| 0.0043 |
47.0 |
5287 |
0.5743 |
0.88 |
| 0.0043 |
48.0 |
5400 |
0.5757 |
0.87 |
| 0.0037 |
49.0 |
5512 |
0.5710 |
0.88 |
| 0.0036 |
49.78 |
5600 |
0.5759 |
0.87 |
框架版本
- Transformers 4.30.2
- Pytorch 2.0.1+cu117
- Datasets 2.13.1
- Tokenizers 0.13.3
📄 許可證
本項目採用Apache-2.0許可證。
相關信息表格
| 屬性 |
詳情 |
| 模型類型 |
微調版DistilHuBERT模型 |
| 訓練數據 |
marsyas/gtzan數據集 |
| 評估指標 |
準確率 |
| 基礎模型 |
ntu-spml/distilhubert |
%20--%3e%3cdefs%3e%3cstyle%3e%20.st0%20{%20fill:%20%23061b40;%20}%20.st1%20{%20fill:%20%23306af1;%20}%20.st2%20{%20fill:%20%235ce5cf;%20}%20%3c/style%3e%3c/defs%3e%3cg%3e%3cpath%20class='st0'%20d='M55,10.5h9v3h-9v9h12V7.5h-12v3ZM64,19.5h-6v-3h6v3Z'/%3e%3cpolygon%20class='st0'%20points='69%2016.5%2078%2016.5%2078%2019.5%2069%2019.5%2069%2022.5%2081%2022.5%2081%2013.5%2072%2013.5%2072%2010.5%2081%2010.5%2081%207.5%2069%207.5%2069%2016.5'/%3e%3cpolygon%20class='st0'%20points='95%2010.5%2095%207.5%2083%207.5%2083%2022.5%2095%2022.5%2095%2019.5%2086%2019.5%2086%2016.5%2095%2016.5%2095%2013.5%2086%2013.5%2086%2010.5%2095%2010.5'/%3e%3cpath%20class='st0'%20d='M40,1.5v21h11.6l1.4-1.4v-7.6h0c0,0-1.4-1.5-1.4-1.5l1.4-1.4V2.9l-1.4-1.4h-11.6ZM50,19.5h-7v-6h7v6ZM50,10.5h-7v-6h7v6Z'/%3e%3c/g%3e%3cpath%20class='st1'%20d='M23.1,24L14.7,4.8l-4.9,11.2h3.8l-1.8,4H3.3L12.1,0H2C.9,0,0,.9,0,2v20c0,1.1.9,2,2,2h21.1Z'/%3e%3cpath%20class='st2'%20d='M34,0h-16.8l10.6,24h6.2c1.1,0,2-.9,2-2V2C36,.9,35.1,0,34,0ZM32.5,20h-4V4h4v16Z'/%3e%3c/svg%3e)
PyTorch
Transformers 支持多種語言