File size: 2,922 Bytes
cc426c6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
6054c7b
cc426c6
 
 
 
 
 
 
 
 
6054c7b
 
cc426c6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
---
license: apache-2.0
base_model: bookbot/distil-ast-audioset
tags:
- generated_from_trainer
datasets:
- marsyas/gtzan
metrics:
- accuracy
model-index:
- name: distil-ast-audioset-finetuned-gtzan
  results:
  - task:
      name: Audio Classification
      type: audio-classification
    dataset:
      name: GTZAN
      type: marsyas/gtzan
      config: train
      split: train
      args: train
    metrics:
    - name: Accuracy
      type: accuracy
      value: 0.92
---

<!-- 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. -->

# distil-ast-audioset-finetuned-gtzan

This model is a fine-tuned version of [bookbot/distil-ast-audioset](https://huggingface.co/bookbot/distil-ast-audioset) on the GTZAN dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3742
- Accuracy: 0.92

## 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

### Training results

| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| 1.0362        | 1.0   | 113  | 0.8029          | 0.75     |
| 0.7068        | 2.0   | 226  | 0.4494          | 0.86     |
| 0.3458        | 3.0   | 339  | 0.5926          | 0.78     |
| 0.1416        | 4.0   | 452  | 0.5162          | 0.89     |
| 0.1134        | 5.0   | 565  | 0.6216          | 0.84     |
| 0.003         | 6.0   | 678  | 0.5221          | 0.9      |
| 0.3177        | 7.0   | 791  | 0.3742          | 0.92     |
| 0.0002        | 8.0   | 904  | 0.5134          | 0.91     |
| 0.0022        | 9.0   | 1017 | 0.6814          | 0.86     |
| 0.0002        | 10.0  | 1130 | 0.5605          | 0.9      |
| 0.0001        | 11.0  | 1243 | 0.4976          | 0.9      |
| 0.0105        | 12.0  | 1356 | 0.5841          | 0.89     |
| 0.0001        | 13.0  | 1469 | 0.4664          | 0.89     |
| 0.0           | 14.0  | 1582 | 0.5755          | 0.89     |
| 0.0           | 15.0  | 1695 | 0.4913          | 0.9      |
| 0.0001        | 16.0  | 1808 | 0.4877          | 0.9      |
| 0.0           | 17.0  | 1921 | 0.4855          | 0.9      |
| 0.0           | 18.0  | 2034 | 0.4859          | 0.9      |
| 0.0           | 19.0  | 2147 | 0.4862          | 0.9      |
| 0.0           | 20.0  | 2260 | 0.4865          | 0.9      |


### Framework versions

- Transformers 4.32.0.dev0
- Pytorch 2.0.1+cu118
- Datasets 2.14.4
- Tokenizers 0.13.3