Skip to content

This is a code implementation of the paper Invariant Information Clustering for Unsupervised Image Classification and Segmentation

License

Notifications You must be signed in to change notification settings

stefanherdy/IIC

Repository files navigation

unsupervised-invariant-information-clustering

This is a code implementation of the paper Invariant Information Clustering for Unsupervised Image Classification and Segmentation.

Installation

$ git clone https://github.com/stefanherdy/unsupervised-invariant-information-clustering.git

Usage

- First, add your custom datasets to the input_data folder
- Run sem_seg_test.py
    You can specify the following parameters:
    --learnrate, type=int, default=0.001, help='learn rate of optimizer"
    --epochs, type=int, default=500
    --batch_size, type=int, default=2, help="Batch Size"
    --num_layers, type=int, default=32, help="Number of UNet layers"
    --num_blocks, type=int, default=1, help="Number of UNet blocks"
    --resize, type=int, default=512, help="Image size for resizing"

    Example usage:
    "python train.py --batch_size 4 --learnrate 0.0001 --resize 1024

- optimize your hyperparameters

License

This project is licensed under the MIT License. ©️ 2023 Stefan Herdy

About

This is a code implementation of the paper Invariant Information Clustering for Unsupervised Image Classification and Segmentation

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages