File size: 2,717 Bytes
8884f5c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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: ntu-spml/distilhubert
tags:
- generated_from_trainer
datasets:
- Emo-Codec/CREMA-D_synth
metrics:
- accuracy
- precision
- recall
- f1
model-index:
- name: distilhubert-tone-classification
  results:
  - task:
      name: Audio Classification
      type: audio-classification
    dataset:
      name: CREMA-D
      type: Emo-Codec/CREMA-D_synth
    metrics:
    - name: Accuracy
      type: accuracy
      value: 0.6809651474530831
    - name: Precision
      type: precision
      value: 0.6795129218164245
    - name: Recall
      type: recall
      value: 0.6809651474530831
    - name: F1
      type: f1
      value: 0.6750238551197275
---

<!-- 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. -->

# distilhubert-tone-classification

This model is a fine-tuned version of [ntu-spml/distilhubert](https://huggingface.co/ntu-spml/distilhubert) on the CREMA-D dataset.
It achieves the following results on the evaluation set:
- Loss: 1.1796
- Accuracy: 0.6810
- Precision: 0.6795
- Recall: 0.6810
- F1: 0.6750

## 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: 16
- eval_batch_size: 16
- 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: 8

### Training results

| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1     |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:---------:|:------:|:------:|
| 1.3122        | 1.0   | 442  | 1.1656          | 0.5737   | 0.5887    | 0.5737 | 0.5679 |
| 1.0131        | 2.0   | 884  | 0.9625          | 0.6461   | 0.6572    | 0.6461 | 0.6399 |
| 0.7817        | 3.0   | 1326 | 1.0005          | 0.6381   | 0.6506    | 0.6381 | 0.6249 |
| 0.6087        | 4.0   | 1768 | 0.9428          | 0.6649   | 0.6572    | 0.6649 | 0.6515 |
| 0.4604        | 5.0   | 2210 | 1.0250          | 0.6622   | 0.6710    | 0.6622 | 0.6545 |
| 0.3164        | 6.0   | 2652 | 1.0814          | 0.6783   | 0.6821    | 0.6783 | 0.6656 |
| 0.2127        | 7.0   | 3094 | 1.1286          | 0.6971   | 0.6991    | 0.6971 | 0.6909 |
| 0.1224        | 8.0   | 3536 | 1.1796          | 0.6810   | 0.6795    | 0.6810 | 0.6750 |


### Framework versions

- Transformers 4.42.4
- Pytorch 2.3.1+cu121
- Datasets 2.21.0
- Tokenizers 0.19.1