File size: 4,498 Bytes
8995214
 
 
 
 
 
 
 
542cd52
 
 
8995214
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
542cd52
 
 
8995214
 
1535a62
8995214
542cd52
 
8995214
 
 
 
 
 
 
 
 
 
 
 
 
1535a62
 
 
 
632f9fe
8d8f1ba
 
1535a62
 
 
8d8f1ba
1535a62
 
 
8d8f1ba
8995214
8d8f1ba
1535a62
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
6c7b19c
 
8995214
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
542cd52
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
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
---
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- imagefolder
metrics:
- accuracy
- f1
- recall
- precision
model-index:
- name: deit-base-distilled-patch16-224-Brain_Tumors_Image_Classification
  results:
  - task:
      name: Image Classification
      type: image-classification
    dataset:
      name: imagefolder
      type: imagefolder
      config: default
      split: train
      args: default
    metrics:
    - name: Accuracy
      type: accuracy
      value: 0.8045685279187818
language:
- en
pipeline_tag: image-classification
---

<h1>deit-base-distilled-patch16-224-Brain_Tumors_Image_Classification</h1>

This model is a fine-tuned version of [facebook/deit-base-distilled-patch16-224](https://huggingface.co/facebook/deit-base-distilled-patch16-224).

It achieves the following results on the evaluation set:
- Loss: 1.8587
- Accuracy: 0.8046
- Weighted f1: 0.7749
- Micro f1: 0.8046
- Macro f1: 0.7814
- Weighted recall: 0.8046
- Micro recall: 0.8046
- Macro recall: 0.7996
- Weighted precision: 0.8567
- Micro precision: 0.8046
- Macro precision: 0.8710

<div style="text-align: center;">
<h2>
    Model Description
</h2>
<a href="https://github.com/DunnBC22/Vision_Audio_and_Multimodal_Projects/blob/main/Computer%20Vision/Image%20Classification/Multiclass%20Classification/Brain%20Tumors%20Image%20Classification%20Comparison/DeiT%20-%20Image%20Classification.ipynb">
    Click here for the code that I used to create this model
</a>
This project is part of a comparison of seventeen (17) transformers. 
<a href="https://github.com/DunnBC22/Vision_Audio_and_Multimodal_Projects/blob/main/Computer%20Vision/Image%20Classification/Multiclass%20Classification/Brain%20Tumors%20Image%20Classification%20Comparison/README.md">
    Click here to see the README markdown file for the full project
</a>
<h2>
    Intended Uses & Limitations
</h2>
This model is intended to demonstrate my ability to solve a complex problem using technology.


<h2>
    Training & Evaluation Data
</h2>
<a href="https://www.kaggle.com/datasets/sartajbhuvaji/brain-tumor-classification-mri">
    Brain Tumor Image Classification Dataset
</a>
<h2>
    Sample Images
</h2>
<img src="https://github.com/DunnBC22/Vision_Audio_and_Multimodal_Projects/raw/main/Computer%20Vision/Image%20Classification/Multiclass%20Classification/Brain%20Tumors%20Image%20Classification%20Comparison/Images/Sample%20Images.png" />
<h2>
    Class Distribution of Training Dataset
</h2>
<img src="https://github.com/DunnBC22/Vision_Audio_and_Multimodal_Projects/raw/main/Computer%20Vision/Image%20Classification/Multiclass%20Classification/Brain%20Tumors%20Image%20Classification%20Comparison/Images/Class%20Distribution%20-%20Training%20Dataset.png"/>
<h2>
    Class Distribution of Evaluation Dataset
</h2>
<img src="https://github.com/DunnBC22/Vision_Audio_and_Multimodal_Projects/raw/main/Computer%20Vision/Image%20Classification/Multiclass%20Classification/Brain%20Tumors%20Image%20Classification%20Comparison/Images/Class%20Distribution%20-%20Testing%20Dataset.png"/>
</div>

## Training procedure

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 0.0002
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3

### Training results

| Training Loss | Epoch | Step | Validation Loss | Accuracy | Weighted f1 | Micro f1 | Macro f1 | Weighted recall | Micro recall | Macro recall | Weighted precision | Micro precision | Macro precision |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:-----------:|:--------:|:--------:|:---------------:|:------------:|:------------:|:------------------:|:---------------:|:---------------:|
| 1.6561        | 1.0   | 180  | 1.5974          | 0.7792   | 0.7454      | 0.7792   | 0.7524   | 0.7792          | 0.7792       | 0.7722       | 0.8318             | 0.7792          | 0.8488          |
| 1.6561        | 2.0   | 360  | 1.7614          | 0.7944   | 0.7575      | 0.7944   | 0.7633   | 0.7944          | 0.7944       | 0.7896       | 0.8458             | 0.7944          | 0.8582          |
| 0.172         | 3.0   | 540  | 1.8587          | 0.8046   | 0.7749      | 0.8046   | 0.7814   | 0.8046          | 0.8046       | 0.7996       | 0.8567             | 0.8046          | 0.8710          |

### Framework versions

- Transformers 4.28.1
- Pytorch 2.0.0
- Datasets 2.11.0
- Tokenizers 0.13.3