Khurram Javed's repositories
incremental-learning
Pytorch implementation of ACCV18 paper "Revisiting Distillation and Incremental Classifier Learning."
real-time-networks
Distributed and Dynamic Neural Network for Generate and Test based Representation Learning
baby-transfomers
A repo for efficiently experimenting with small transformer models that can run efficiently on a single CPU core (For academic purposes).
Columnar-Constructive-Networks
Code for the JMLR paper "Scalable Real-Time Recurrent Learning Using Columnar-Constructive Networks"
real-time-rl-environments
Async version of atari learning environment.
sparse_nn_gpu_benchmark
Code to benchmark nodes on Cedar, Narval, Beluga, and Graham clusters.
Atari-on-the-web
The Arcade Learning Environment (ALE) -- a platform for AI research.
Efficient-softmax-classifier
An efficient implementation of neural network with softmax classifier. Can be used as actors in actor critic algorithm. We use it for MNIST classification in this repo
limitations-of-adam
Code for experiments comparing adam vs meta-descent for step-size adaptation for a linear learner
step-size-adaptation
A simple comparison to demonstrate limitation of adam optimizer for non-stationary problems
The-big-world-hypothesis
Code for the paper on the big world hypothesis