itsayushthada / FastAI2019

This repository created to practice fastai module as well as to benchmark the dataset for various tasks using it. Though I'm following Jeremy Howard's session for now, but I've some other plans for it.

FastAI2019

All the tasks have been done with the FastAI and Pytorch module.

DataSet	Accuracy	Model Name	Task
MNIST	99.40%	ResNet18	Classification
CIFAR100	82.03%	ResNet50	Classification
Devanagari Handwritten Character Dataset	99.1624%	ResNet18	Classification
Arabic Hand Written Dataset	NaN%	ResNet18	Classification

These are not the final results. There is still scope of improvement.