Skip to content

Commit 96105fb

Browse files
committed
intial_readme
1 parent 486c341 commit 96105fb

12 files changed

+11698
-11600
lines changed

.gitignore

+25-25
Original file line numberDiff line numberDiff line change
@@ -1,26 +1,26 @@
1-
2-
### VisualStudioCode ###
3-
.vscode/*
4-
!.vscode/settings.json
5-
!.vscode/tasks.json
6-
!.vscode/launch.json
7-
!.vscode/extensions.json
8-
!.vscode/*.code-snippets
9-
10-
# Local History for Visual Studio Code
11-
.history/
12-
13-
# Built Visual Studio Code Extensions
14-
*.vsix
15-
16-
### VisualStudioCode Patch ###
17-
# Ignore all local history of files
18-
.history
19-
.ionide
20-
21-
*.png
22-
*.zip
23-
DS_10283_3192/
24-
*.mp4
25-
*.pth
1+
2+
### VisualStudioCode ###
3+
.vscode/*
4+
!.vscode/settings.json
5+
!.vscode/tasks.json
6+
!.vscode/launch.json
7+
!.vscode/extensions.json
8+
!.vscode/*.code-snippets
9+
10+
# Local History for Visual Studio Code
11+
.history/
12+
13+
# Built Visual Studio Code Extensions
14+
*.vsix
15+
16+
### VisualStudioCode Patch ###
17+
# Ignore all local history of files
18+
.history
19+
.ionide
20+
21+
*.png
22+
*.zip
23+
DS_10283_3192/
24+
*.mp4
25+
*.pth
2626
Models-Keras/

LICENSE.md

+21
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,21 @@
1+
MIT License
2+
3+
Copyright (c) [year] [fullname]
4+
5+
Permission is hereby granted, free of charge, to any person obtaining a copy
6+
of this software and associated documentation files (the "Software"), to deal
7+
in the Software without restriction, including without limitation the rights
8+
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
9+
copies of the Software, and to permit persons to whom the Software is
10+
furnished to do so, subject to the following conditions:
11+
12+
The above copyright notice and this permission notice shall be included in all
13+
copies or substantial portions of the Software.
14+
15+
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
16+
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
17+
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
18+
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
19+
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
20+
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
21+
SOFTWARE.

README.md

+77
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,77 @@
1+
2+
<!-- PROJECT -->
3+
# Image Classification with Transfer Learning
4+
5+
<!-- ABOUT THE PROJECT -->
6+
7+
## About The Project
8+
This repository, "ImgClassTF" focuses on exploring the effectiveness of transfer learning in image classification tasks. By leveraging pre-trained models, specifically VGG networks, the goal was to utilize pretrained ConvNet models using transfer learning, which significantly improved the classification accuracy with minimal computational resources compared to training models from scratch. Though the accuracy achieved was 53.09%
9+
10+
In my study, I used the VGG model with pretrained weights on [IMAGENET](https://www.image-net.org) and used a modified version of it to train on [CINIC-10](https://arxiv.org/abs/1810.03505). I experimented with various configurations of the VGG model to classify images across multiple categories. The findings, detailed in my research [paper]() and [code](), highlight the power of transfer learning in achieving high accuracy rates in image classification tasks.
11+
12+
Below are some of the results from the paper.
13+
<p align="center">
14+
<img src="RP_Docs/model_4_combined_training_curves.png" alt="Training curves" width="80%" />
15+
</p>
16+
17+
<!-- OVERVIEW -->
18+
## Overview
19+
<br> Transfer Learning Efficiency: Utilizing VGG pre-trained models allows for considerable improvements in accuracy.
20+
<br> Model Comparison: Analysis of different VGG configurations to determine the most effective setup for our classification task.
21+
<br> Performance Metrics: Evaluation of models based on accuracy, precision, and recall metrics.
22+
23+
<!-- FLOWCHART -->
24+
## General Flowchart for the project
25+
<p align="center">
26+
<img src="RP_Docs/CNN.png" width="400">
27+
</p>
28+
29+
<!-- REIMPLEMENTATION -->
30+
## Implementing
31+
The project is implemented using a jupyter notebook, so its fairly straightforward to download it directly to retest. [Notebook]()
32+
I used 4 versions of the models
33+
1. VGG16 model with untrainable Convolutional layers and replacing the classifier with a flatten layer, dense layer with Relu activation and output layer with Softmax activation function
34+
<img src="RP_Docs/m1.png" width="400">
35+
36+
2. The configuration is the same as model 1, in addition to that we introduce data augmentation for training the model
37+
<img src="RP_Docs/m2.png" width="400">
38+
39+
3. The configuration is the same as model 1, in addition to that we introduce data augmentation for training and regularization methods such as Batch Normalization and Dropout Layer
40+
<img src="RP_Docs/m3.png" width="400">
41+
42+
4. VGG16 model with fully trainable weights and replacing the classifier with a flatten layer, dense layer with Relu activation and output layer with Softmax activation function.
43+
<img src="RP_Docs/m4.png" width="400">
44+
45+
46+
<!-- Dependencies -->
47+
## Dependencies
48+
49+
The code utilizes the python packages as such
50+
* matplotlib,
51+
* VisualKeras,
52+
* pickle,
53+
* numpy,
54+
* scikit-learn,
55+
* Tensorflow.
56+
57+
<!-- LICENSE -->
58+
## License
59+
Distributed under the MIT License. See `LICENSE` for more information.
60+
61+
<!-- Authors -->
62+
## Authors
63+
Project Link:[Project]()
64+
<br>~~Kaggle Notebook:[]()~~
65+
<br>~~Collab Notebook:[]()~~
66+
67+
## Thank you
68+
69+
<picture>
70+
<source media="(prefers-color-scheme: dark)" srcset="https://user-images.githubusercontent.com/25423296/163456776-7f95b81a-f1ed-45f7-b7ab-8fa810d529fa.png">
71+
<source media="(prefers-color-scheme: light)" srcset="https://user-images.githubusercontent.com/25423296/163456779-a8556205-d0a5-45e2-ac17-42d089e3c3f8.png">
72+
<img alt="Shows an illustrated sun in light mode and a moon with stars in dark mode." src="https://user-images.githubusercontent.com/25423296/163456779-a8556205-d0a5-45e2-ac17-42d089e3c3f8.png">
73+
</picture>
74+
75+
> [!NOTE]
76+
> This is part of the Assignment of COMP SCI 7318 - Deep Learning Fundamentals, University of adelaide, ~~which had the worst Teaching assistants imo :vomiting_face:~~.
77+
> I did not conduct hyperparameter tuning of the model, which is an essential step for achieving good accuracy.

0 commit comments

Comments
 (0)