https://github.com/andrew2077/102flowers-eda-classification

FellowshipAi project
https://github.com/andrew2077/102flowers-eda-classification

explanatory-data-analysis flowers102 gradcam-visualization python pytorch tensorboard xai

Last synced: about 2 months ago
JSON representation

FellowshipAi project

• Host: GitHub
• URL: https://github.com/andrew2077/102flowers-eda-classification
• Owner: Andrew2077
• Created: 2023-08-30T07:31:45.000Z (over 1 year ago)
• Default Branch: main
• Last Pushed: 2023-09-07T07:39:42.000Z (over 1 year ago)
• Last Synced: 2023-09-07T08:58:06.620Z (over 1 year ago)
• Topics: explanatory-data-analysis, flowers102, gradcam-visualization, python, pytorch, tensorboard, xai
• Language: Python
• Homepage:
• Size: 39.7 MB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Flowers102-Resnet50

FellowshipAi project 

might takes some time to load, due to many plots and gifs, please be patient

[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/drive/1M_odF1YhijOgr3FnrRSEtlDMCC7QRgQi?usp=sharing)

The Notebook has everything from downloading dataset to fine tuning the model and inference.

## Preprocessing

- [download.py](engine/data_download.py)

- [data_processing.py](engine/data_processing.py)

### Dataset

- [Flowers102](https://www.robots.ox.ac.uk/~vgg/data/flowers/102/index.html)

### Augmentation

- RandomResizedCrop

- RandomHorizontalFlip

- RandomVerticalFlip

- randomRotation

- CenterCrop

## Trnasformations

  - resize = 224

  - ToTensor

  - Normalize

    - mean = [0.485, 0.456, 0.406]

    - std = [0.229, 0.224, 0.225]

## Model 

- [model.py](engine/model.py)

### Architecture

- Resnet50 - backbone

- MlpHead - head

  - 3 layers

    - 2048 -> 512 -> 102 hidden units

    - ReLU activation

    - Dropout 0.5

    - BatchNorm

    - 

### Train 

  - CrossEntropyLoss

  - Adam optimizer

    - 0.01 learning rate

  - 30 epochs

### Fine-Tuning

  - CrossEntropyLoss

  - Radam optimizer

    - 3e-4 learning rate

  - Cyclical Learning Rates

    - min_lr = 5e-8

    - max_lr = 3e-3

    - mode = 'triangular'

  - 120 epochs

## Results 

### 1. EDA 

#### Classest Distribution

- Class imbalance is present in the dataset

![class distribution](misc/class_dist.png)

#### Sample Images

![sample images](misc/default_sample.gif)

### Training History

##### Highest Accuracy

- Test-Set

  

| Experiment | Accuracy | Loss |

| :---: | :---: | :---: |

| Resnet50 - 30 Epochs | 85.0838% | 0.549521 |

| Resnet50 - 120 Epochs | 89.2956% | 0.386665 |

- Validation-Set

| Experiment | Accuracy | Loss |

| :---: | :---: | :---: |

| Resnet50 - 30 Epochs | 87.2511% | 0.451927 |

| Resnet50 - 120 Epochs | 91.0849% | 0.312610 |

- Train-Set

| Experiment | Accuracy | Loss |

| :---: | :---: | :---: |

| Resnet50 - 30 Epochs | 88.8839% | 0.389938 |

| Resnet50 - 120 Epochs | 96.7953% | 0.105484 |

### Graphs

##### LOSS 

- 30 epochs

![30 Epochs Loss](misc/TF_loss.png)

- 120 epochs (fine tuning)

![120 Epochs Loss](misc/Fine_tune_loss.png)

##### ACCURACY

- 30 epochs

![30 Epochs Accuracy](misc/TF_acc.png)

- 120 epochs (fine tuning)

![120 Epochs Accuracy](misc/Fine_tune_acc.png)

### GradCam

#### Train Sample 

- grad camp for 1 image & top 4 classes

![gradcam](misc/sample_train.png)

- CAM for all classes

![gradcam](misc/all_feat_train.gif)

#### Val Sample 

- idx 75

- grad camp for 1 image & top 4 classes

![gradcam](misc/sample_val.png)

- CAM for all classes

![gradcam](misc/all_feat_val.gif)

#### Test Sample 

- idx 75

- grad camp for 1 image & top 4 classes

![gradcam](misc/sample_test.png)

- CAM for all classes

![gradcam](misc/all_feat_test.gif)

## [Workflow](workflow.md)