File size: 2,040 Bytes
69160ee
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
license: bsd-3-clause
tags:
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: ast_9-finetuned-ICBHI
  results: []
---

<!-- 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. -->

# ast_9-finetuned-ICBHI

This model is a fine-tuned version of [MIT/ast-finetuned-audioset-10-10-0.4593](https://huggingface.co/MIT/ast-finetuned-audioset-10-10-0.4593) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.8548
- Accuracy: 0.6572
- Sensitivity: 0.4183
- Specificity: 0.8710
- Score: 0.6446

## 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: 3e-05
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 5

### Training results

| Training Loss | Epoch | Step | Validation Loss | Accuracy | Sensitivity | Specificity | Score  |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:-----------:|:-----------:|:------:|
| 0.9335        | 1.0   | 258  | 0.9986          | 0.5986   | 0.5318      | 0.6582      | 0.5950 |
| 0.8392        | 2.0   | 517  | 0.8548          | 0.6572   | 0.4183      | 0.8710      | 0.6446 |
| 0.6187        | 3.0   | 776  | 0.8978          | 0.6554   | 0.4797      | 0.8126      | 0.6461 |
| 0.2052        | 4.0   | 1035 | 1.1767          | 0.6533   | 0.4536      | 0.8318      | 0.6427 |
| 0.1132        | 4.99  | 1290 | 1.4218          | 0.6525   | 0.5004      | 0.7886      | 0.6445 |


### Framework versions

- Transformers 4.29.1
- Pytorch 2.0.0+cu118
- Datasets 2.12.0
- Tokenizers 0.13.3