✋🏼🛑 This one stop project is a complete COVID-19 detection package comprising of 3 tasks: • Task 1 --> COVID-19 Classification • Task 2 --> COVID-19 Infection Segmentation • Task 3 --> Lung Segmentation

Keras callbacks for one-cycle training, cyclic learning rate (CLR) training, and learning rate range test.

This repository contains the jupyter notebooks for the custom-built DenseNet Model build on Tiny ImageNet dataset

training food-101 (achieved SOTA top-1 validation acc ~=90%) using 1-cycle-policy:

Cyclical Learning Rate and 1Cycle Policy as Keras callback.

As the learning rate is one of the most important hyper-parameters to tune for training convolutional neural networks. In this paper, a powerful technique to select a range of learning rates for a neural network that named cyclical learning rate was implemented with two different skewness degrees. It is an approach to adjust where the value is c…

Paper to Code automates the incorporation of research paper concepts into practical code using OpenAI's GPT models, bridging theory and implementation.

Pytorch implementation of the paper: 'Cyclical Learning Rates for Training Neural Networks'

This repository contains the Jupyter notebook for the custom-built VGG16 Model build for the Tiny ImageNet dataset.

This repository contains the Jupyter notebook for the custom-built VGG16 Model build for the Tiny ImageNet dataset.

Using the pre-trained ImageNet models and cyclical learning rates, we tried to classify the DeepSAT-6 dataset (https://csc.lsu.edu/~saikat/deepsat/) into 6 categories (barren land, trees, grassland, roads, buildings and water bodies). Due to the absence of occlusion by the cloud, we dropped the NIR channel of the data.

Classify footware based on closures : https://nbviewer.jupyter.org/github/shubhajitml/footware/tree/master/

self-used pytorch utilities

Deep Learning for Insincere Question Classification