Added README.md and results to share

README.md

@@ -0,0 +1,21 @@
+# Deep Convolution Generative Adversarial Networks
+
+This repository is an implementation of DCGAN's that utilize convolution layers in their network architecture to produce images and learn mappings from 1x100 dimenssional latent space to target images. The power of these networks to learn connections within this latent space is also admirable.
+
+We ran this model on three datasets:
+
+1. CelebA Dataset: http://mmlab.ie.cuhk.edu.hk/projects/CelebA.html
+2. Simpsons Dataset: https://www.kaggle.com/kostastokis/simpsons-faces
+3. Simplified Simpsons Dataset: https://www.kaggle.com/kostastokis/simpsons-faces
+
+To make the training easier, we downsized the images to 64x64 for datasets 1 and 3 and 128x128 for dataset 2.
+
+Link to the original paper: https://arxiv.org/abs/1511.06434
+
+## Results
+
+
+<div style="display: flex; flex-direction: column; align-items: center">
+<img src="./celeba_sample.png" alt="Celebrity images generated my model" style="width: 400px">
+After 50000 epochs, the samples from CelebA dataset
+</div>

dir.sh

@@ -0,0 +1,28 @@
+#!/bin/bash
+
+#used to make the necessary directories to store results from experiments
+mkdir samples
+mkdir plots
+mkdir models
+mkdir gifs
+mkdir figures
+
+mkdir samples/celeba_64
+mkdir samples/simpsons_128
+mkdir samples/simpsons_simplified_64
+
+mkdir figure/celeba_64
+mkdir figure/simpsons_128
+mkdir figure/simpsons_simplified_64
+
+mkdir models/celeba_64
+mkdir models/simpsons_128
+mkdir models/simpsons_simplified_64
+
+mkdir gifs/celeba_64
+mkdir gifs/simpsons_128
+mkdir gifs/simpsons_simplified_64
+
+mkdir plots/celeba_64
+mkdir plots/simpsons_128
+mkdir plots/simpsons_simplified_64

requirements.txt

@@ -0,0 +1,15 @@
+cycler==0.11.0
+dataclasses==0.8
+imageio==2.13.5
+kiwisolver==1.3.1
+matplotlib==3.3.4
+numpy==1.19.5
+opencv-python==4.5.5.62
+Pillow==8.4.0
+pyparsing==3.0.6
+python-dateutil==2.8.2
+six==1.16.0
+torch==1.10.1
+torchvision==0.11.2
+tqdm==4.62.3
+typing_extensions==4.0.1