File size: 1,922 Bytes
1a9d935
 
 
 
 
 
 
 
 
 
 
f8a97e1
 
1a9d935
 
 
 
 
 
 
 
 
 
a1aea59
1a9d935
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
a1aea59
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
---
library_name: transformers
license: bsd-3-clause
base_model: MIT/ast-finetuned-audioset-10-10-0.4593
tags:
- generated_from_trainer
metrics:
- accuracy
- recall
- precision
- f1
language:
- en
model-index:
- name: AST_EmoRecog_Model_v3
  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_EmoRecog_Model_v3

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 [IEMOCAP](https://sail.usc.edu/iemocap/) dataset.
It achieves the following results on the evaluation set:
- Loss: 1.2233
- Accuracy: 0.5299
- Recall: 0.3967
- Precision: 0.5460
- F1: 0.3992

## 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: 1e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- num_epochs: 3

### Training results

| Training Loss | Epoch | Step | Validation Loss | Accuracy | Recall | Precision | F1     |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|:---------:|:------:|
| 1.4346        | 1.0   | 377  | 1.2955          | 0.5186   | 0.3738 | 0.5319    | 0.3590 |
| 1.138         | 2.0   | 754  | 1.2554          | 0.5212   | 0.3942 | 0.5089    | 0.3824 |
| 0.9068        | 3.0   | 1131 | 1.2233          | 0.5299   | 0.3967 | 0.5460    | 0.3992 |


### Framework versions

- Transformers 4.47.1
- Pytorch 2.5.1+cu121
- Datasets 3.2.0
- Tokenizers 0.21.0