Skip to content

Latest commit

 

History

History
50 lines (33 loc) · 1.34 KB

README.md

File metadata and controls

50 lines (33 loc) · 1.34 KB

SCL-IKD: intermediate knowledge distillation via supervised contrastive representation learning

Published at Applied intelligence Journal

Here is the link to the article.

Requirements

The code runs correctly with

  • Python 3.7
  • Keras 2.4.0
  • TensorFlow 2.4.0

Files

  • CIFAR_10.py
  • CIFAR_100.py
  • TinyImagenet.py

How to run

# GPU Id's
Set the corresponding GPU's Id's in respective codes.

# Install the basic libraries
Open the code and install all the libraries accordingly in given sequence.

# Create the CONDA Environment
Create the new conda environment ot run the files

# Run the file
All things are setup and just activate the conda environment and run python Filename.py for running the desired file.

Data Preparation

All the dataset have been already imported in the corresponding codes for CIFAR10 and CIFAR100 dataset. In case of Tinyimagenet dataset instruction will be given in the file for downloading the dataset and setting the path of the dataset.

Contact

please contact [email protected] for any discrepancy.

Authors:

  • Saurabh Sharma
  • Shikhar Singh Lodhi
  • Joydeep Chandra

Note

The code provided have the random hyperparameter setting and will provide the basic idea. To reproduce the results, please feel free to contact.