ayush055 Goto Github PK

botui

Robot User Interface for the KIPR Controllers (Link, Wallaby, Wombat)

cassava-leaf-disease-classification

This project was part of the Kaggle Cassava Leaf Disease Classification competition. This was an image classification competition in which I built a machine learning model capable of classifying images of Cassava leaves into one of five different categories. This project used transfer learning, fine-tuning an EfficientNetB3.

cots-object-detection

This project was part of the Kaggle COTS Object Detection competition. This was an object detection competition in which I used a YOLOv5 model to identify the location of COTS in video data.

hurricane-trajectory-prediction

This was a project that used data from the SHIPS dataset to predict hurricane trajectories. An LSTM Model was built in PyTorch that can predict the trajectories of hurricanes given surrounding environmental conditions. Different techniques such as teacher forcing were used to train the model.

libwallaby

Library for interfacing with the Wombat firmware

lung-cancer-research

This project was conducted using data from the LIDC-IDRI dataset. Two 3D CNN models were built, one for classification of lung nodules and another for the segmentation of lung nodules from CT scans.

mood-tracker

mood-tracker-ml

open-unmix-pytorch

Open-Unmix - Music Source Separation for PyTorch

signslate

Signslate is an application that allows users to speak and translate speech to ASL or allow users to sign in ASL through their webcam and translate it to speech.

trie-project

Py-Trie is a command-line interface which enables communication with a global trie hosted on the Google Cloud Platform. Users can execute basic commands such as inserting, deleting, and searching words in the trie. A list of matching words can be generated through an input prefix. Users are also able to display the words present in the trie.