venkateshtata Goto Github PK

-convolutional-neural-network

This is a convolutional neural network trained with more than 15,000 dogs and cat images respectively. I have built this network three layers deep, giving an accuracy of 86%.

algorithms_in_cpp

anonymous-chat-room

This a chat room application created using Node.js, socket.io, and html.

bank_costumer_prediction

This is an artificial neural network that is trained with 10'000 people details who were or are the costumers of the respective bank. This ANN is going to take the required details of a costumer and predict if e is going to leave the bank or not. This model is totally made using the Keras library by running tensorflow backend and Numpy for numerical computations.

blog

brainweight

A simple tensorFlow script for visualising the dataset in graph using Pandas.

calibrated_pointcloud

chat_example

cnn_medium.

A simple binary classifier to predict if the given image is a cat or a dog.

compilergym

A reinforcement learning toolkit for compiler optimizations

computer_vision_security_system

For Vistara Aviation Hackathon

computervision_with_cpp

C++ implementation of computer vision operations

controller

denoise

diffusion_with_lora

disparity_map

distillbert_quantized

drive_autonomous

Objective : To make the concept of autonomous driving cars a real thing by providing cheap affordable and intelligent hardware upgrade that can convert a car to Level 4 autonomous car, to the people without the need for them to upgrade to a total new car. * Our Idea is to Build a portable Artificial INtelligence system that can talk to your car drive the car autonomously at Level 4, this device/system can be installed in particular variety of Benz Cars that have supporting hardware allowance. So this is how we are planning to build it : 1. We are going to first build a deep learning based Neural Network model containing convolutional and recurrent layers. 2. We are planning to train the above model using a virtual environment such as GTA 5 game, as it has more realistic game physics (For the Hackathon purpose, the same model can also be later trained in real-world environment). 3. The above neural network model will be hosted onto a raspberry pi, which has few slave microcontrollers that are going to mimic the real-world car dynamics. 4. Once our Artificial Intelligence Model is giving us proper/required accuracy, we are going to feed the visual data of a road scenario to the camera connected to our A.I's Brain(i.e Raspberry pi), then we can see the behaviour of the microcontrollers and the servos attached to them react/change according to what the camera is seeing.

ea

facebook-newsbot

A facebook messenger chatbot who's server lies on a raspberry-pi and supports NLP(Natural Language Processing) using wit.ai to serve the user while chatting.

facenet

Implementation of FaceNet.

fashion-vector

fav_movies

A python based web api to display your favourite movies.

flappernews

A reddit clone using Angular.js, Node.js and other js frameworks

flippingcoin

freelance_work

furore

Its basically a simple registration app for android, a very good one to refer for those who are building for the first time

gan_medium

This is a Generative Adversarial Network model, where two artificial neural networks compete against each other, result of which will be images birds that are out of this world.

http-server-using-node

idm-vton

IDM-VTON : Improving Diffusion Models for Authentic Virtual Try-on in the Wild