bhaba-ranjan / Image-Models

My Pytorch Implementation of some basic models.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

VGG16

My Pytorch Implementation of VGG16 based off of https://arxiv.org/pdf/1409.1556.pdf

Here is a sample of the classification on CIFAR-10 after about 20 epochs of training:

Alt text

Classifications:

Horse, Dog, Ship, Plane Ship, Deer, Horse, Plane Cat, Horse, Cat, Ship Dog, Dog, Car, Horse

Resnet34

My Pytorch Implementation of Resnet34 from https://arxiv.org/pdf/1512.03385.pdf. This model performs slightly better than VGG16, but with less training time and takes about half as much to classify a new example.

Here is a sample of the classification on CIFAR-10 after about 7 epochs of training:

Alt text

Classifications:

Dog, Car, Ship, Plane, Frog, Frog, Car, Frog, Cat, Car, Plane, Truck, Dog, Horse, Truck, Ship

GAN

My Pytorch implementation of a Generative Adverserial Network is based on https://arxiv.org/pdf/1406.2661.pdf

Below is an example of 25 randonly selected examples from the generator over 15 epochs of training on the MNIST dataset:

Alt text

Autoencoder

Below is an example of 25 randonly selected examples from the generator over 15 epochs of training on the MNIST dataset, also included are the encoding layer which was chosen to be of only size 4:

Alt text Alt text

Logistic Regression

Here is an example of the weights from the logistic and softmax regressions on MNIST, SGD was used to fit them and there were 10000 samples chosen.

Logistic Regression (0/1):

Alt text

Softmax Regression:

Alt text 0 Alt text 1 Alt text 2 Alt text 3 Alt text 4
Alt text 5 Alt text 6 Alt text 7 Alt text 8 Alt text 9

About

My Pytorch Implementation of some basic models.


Languages

Language:Python 100.0%