OpenMMLab Pre-training Toolbox and Benchmark

Facial Landmark Detection in Python, using OpenCV for real-time video capture and basic face detection. This project lays the groundwork for an advanced facial landmark model, utilizing a pre-trained face detection model and a custom-trained CNN for accurate facial landmark identification.

Simplified PyTorch implementation of image classification, support CIFAR10, CIFAR100, MNIST, custom dataset, multi-gpu training and validating, automatic mixed precision training, knowledge distillation, hyperparameter optimization using Optuna etc.

Creating a software for automatic monitoring in online proctoring

MobileNet for Image Classification

Cataract detection model

MobileNetV3 implementation from scratch using PyTorch.

DeepLensX is a Streamlit app that integrates MobileNetV2 and a CIFAR-10 model for image classification. Users can upload images and receive predictions with confidence scores from either model. It features a sleek navigation bar for easy switching and real-time results, ideal for both learning and practical use.

Airplanes Tracking is a system that detects airplanes in sight using MobileNet and then tracks them with a camera, enabling real-time monitoring and analysis.

Segmentation models with pretrained backbones. Keras and TensorFlow Keras.

This repository contains my BSc Thesis materials.

This repository includes multiple exercises about Convolutional Neural Networks

PyTorch implementation of MobileNetV4 family

💎A high level pipeline for face landmarks detection, it supports training, evaluating, exporting, inference(Python/C++) and 100+ data augmentations, can easily install via pip.

Statistical Methods for Machine Learning project - Muffins Vs Chihuahuas

Neural network model repository for highly sparse and sparse-quantized models with matching sparsification recipes

MobileNet V1 implementation from scratch

Support project for my Bachelor's thesis @ Unifi - Converting Oxford5k and Paris6k into datasets suitable for object detection tasks. This repository contains scripts to create XML (PASCAL VOC format) and JSON (COCO format) annotations from .pkl files, enabling the training of an SSD/MobileNet object detection model.

Practice on cifar100(ResNet, DenseNet, VGG, GoogleNet, InceptionV3, InceptionV4, Inception-ResNetv2, Xception, Resnet In Resnet, ResNext,ShuffleNet, ShuffleNetv2, MobileNet, MobileNetv2, SqueezeNet, NasNet, Residual Attention Network, SENet, WideResNet)

Convolutional Neural Networks to predict the aesthetic and technical quality of images.