Mrinal Walia's repositories
AI-for-Finance-Stocks-real-time-analysis-
1. First we fetch data of stocks in realtime from nse India website, perform basis data visualizations using python to analyze the stock. 2. Then we use machine learning LSTM technique to predict the future stock price and at last create an interactive web-app using Streamlit in python.
Facial-Expression-Recognition
• First we recognize the emotion of the person using Opencv and Keras by training our model on Data provided from Kaggle, program is trained for 30 epochs and we have got an accuracy of 71%. • After detecting the emotion out of 7 labels, we as user for his favourite artist and then recommend songs from Spotify using its API and movies from IMDB depending upon its mood.
Face-Classification-into-Blood-No-Blood
Competition
Network-Project-Attacks
Various network attacks
season-of-docs
Supporting materials for Google's Season of Docs 2021
Sitting-Posture-Recognition
Final Year Project
Building-A-Baseline-Convolutional-AutoEncoder-Network-For-Image-Denoising-On-Fashion-MNIST-Dataset
We will build a simple baseline autoencoder model using TensorFlow and the CNN network. Then we will use this network on the FASHION MNIST dataset to show our results and accuracy. We will evaluate our model using a simple CNN network to show how our autoencoder model performs better than a stand-alone CNN model.
COMP8610_Assignment_1
Assignment_1
COMP8610_Assignment_2
Aayushi Navinchandra Patel (Student ID:110087817) Aaditya Pradipbhai Parekh (Student ID:- 110084734) Dhruvkumar Arvind Patel (Student ID:- 110055817) Mrinal Walia (Student ID:- 110066886)
design
This is the main hub for those interested in design in the OpenMined community
EnvisEdge
Deploy recommendation engines with Edge Computing
FederatedScope
An easy-to-use federated learning platform
Generating-Anime-Character-Faces-Using-ACGANs
Abstract-This paper will implement a baseline Auxiliary Classifier Generative Adversarial Network capable of generating new anime character faces. We aim to design a style-guided discriminator and a generator network and train it on 63,000 synthesized anime faces. ACGANs can be notoriously hard to train as the generator, and the discriminator architectures are susceptible to parameters, hyperparameters, regularization, learning rate, and activation functions. However, unlike a conventional GANs network, our proposed framework uses a GPU for faster training and can improve its performance given more c- omputational power and datasets. This problem has previously been solved but lacks a generic architecture. Hence, our proposed framework can be used as a baseline network across different applications of GANs. The generated anime faces from our final model are visually pleasing and resemble designer-generated characters. As generative modelling is an unsupervised task and does not expect images to have labels, the dataset used in this project has 63,000 unlabeled cropped anime faces available at this URL: https://github.com/bchao1/Anime-Face-Dataset. Abbreviations: Auxiliary Classifier Generative Adversarial Networks (ACGANs), Generative Adversarial Networks (GANs), Artificial Intelligence (A.I.), Machine Learning (ML), Convolutional Neural Network (CNN)
geomstats
Computations and statistics on manifolds with geometric structures.
google-images-download
Google/Bing Images Web Downloader
ip-1_assignment-1_backend
Internship/Project-1 Assignment 1 - Master of Applied Computing @ University of Windsor
Kubernetes-Docs-Enhancement
This repository aims to improve the clarity and quality of Kubernetes documentation to make it more accessible and user-friendly for individuals with varying levels of Kubernetes expertise.
Mrinal-Portfolio-Hugo
Hugo: Portfolio
mrinalwalia
-Portfolio Website: Hugo
Open-Source-Requests
A curation of paid and unpaid requests for the community to work on.
PySyft
A library for answering questions using data you cannot see
reverse-interview
Questions to ask the company during your interview
Root-Cause-Analysis-Deep-Learning-Analytis-
IT Operation: Root Cause Analysis
ShootAR
Comp ASE Game.