This project offers an easy, flexible, modular PyTorch implementation for semantic segmentation to minimize configuration, automate training and deployment, and enable customization of models, encoders, losses and datasets through its modular design.

This is a classification model and a post-processing method of any object detection model for false positive elimination.

A framework to accelerate AI/ML development 🚀

Develope a robust and high-performance model utilizing computer vision techniques to classify images as either fraudulent or non-fraudulent within the context of insurance claims

Low-code machine learning and deep learning

Android application used for organizing, sorting and localizing of pictures of landmarks, as well as identifying similar images to the photos made by the user

We convert timm (Pytorch Image Models) models to .onnx and check their performance in browser using ONNX Runtime Web (ort-web). So you can find the suitable model for your JavaScript web-app according to your needs.

Main repo for the project containing the source, sagemaker pipelines, dockers and documentations

A project for MlOps course of ODS

This goal of this repository was to minimize the number of code edits by enabling easy configuration of the Image Classifier pipeline using Hydra, Timm & Lightning.

Simple implementation of CBIR with a pretrained ViT model and local Deeplake DB

Implementation of Barlow Twins: Self-Supervised Learning via Redundancy Reduction with `timm` integration

PyTorch implementation of 'ViViT' (Arnab et al., 2021) from scratch.

Streamlit app that performs binary and multiclass classification of gravitational lensing images along with dark matter halo mass prediction.

An easy classification implement to explain how triton work

PetFinder.my - Pawpularity Contest 13/3537 solution

Implementation of ViT with PyTorch

Image search engine