akshaychawla Goto Github PK

10701-project

GitHub repository containing the code for 10-701 project - Caltech 101 image classification

1d-kalman-filter

This is a simple 1 dimensional Kalman Filter. The Aim of this project was to understand the basics of the Kalman Filter so I could move on to the Extended Kalman Filter.

accelerated-training-by-disentangling-neural-representations

Just a theory.

activation-functions

Testing out different activation functions.

adversarial-examples-in-pytorch

Pytorch code to generate adversarial examples on mnist and ImageNet data.

binary-neural-networks

Exploring "Binary Neural Networks" (https://arxiv.org/abs/1602.02830) in Theano. A set of experiments that use binarised weights and/or activations to reduce computational load of convolutional neural networks.

can-you-hear-me

cnn_finetune

Fine-tune CNN in Keras

code-repos-links

Links to code written in my own and others repositories

core

Cloud Robotics Core: Kubernetes, Federation, App Management

devise-keras

Interpretable Image Search by Priyam Tejaswin and Akshay Chawla

docker_envs

Docker environment data

exploring-cifar-10-keras

google-cloud

hacktoberfest

Participate in Hacktoberfest by contributing to any Open Source project on GitHub! Here is a starter project for first time contributors. #hacktoberfest

image-projection

This is a small python script (APIs included) that warps a source image onto a position in the destination image.

imagenet-downloader

A simple way to browse ImageNet classes and download some images from it.

keras

Deep Learning library for Python. Runs on TensorFlow, Theano, or CNTK.

learning-to-transduce

Learning to transduce with unbounded memory

list-of-publications

This is just a list of publications that I have worked on / am working on.

memless-chan-pool

Memory less channel pooling for TN

neural-network-python-module

A simple and easy to use back-propagation module for python.

occlusion-experiments-for-image-segmentation

Follows the image occlusion pipeline presented in the paper "Visualising and understanding Convolutional Neural networks" to extract a coarse segmentation map from input image using a pre-trained CNN.

optical-flow-tracking----opencv

This is a small program demonstrating object tracking in a video stream. Lucas Kanade optical flow algorithm is used to find the pixels from one frame to another.

pytorch-cifar

95.16% on CIFAR10 with PyTorch

supersonic

A new deep learning framework that GUARANTEES 5x faster training and 8x inference speedups.

vim-configuration

My .vimrc files and vim color schemes

wonderwall

yolov3

YOLOv3 in PyTorch > ONNX > CoreML > iOS