regnet

A tensorflow2 implementation of some basic CNNs(MobileNetV1/V2/V3, EfficientNet, ResNeXt, InceptionV4, InceptionResNetV1/V2, SENet, SqueezeNet, DenseNet, ShuffleNetV2, ResNet).

Pytorch implementation of network design paradigm described in the paper "Designing Network Design Spaces"

Multi-label classification based on timm.

PyTorch-style and human-readable RegNet with a spectrum of pre-trained models

A PyTorch-based Python library with UNet architecture and multiple backbones for Image Semantic Segmentation.

Here is an implementation of DeepLabv3+ in PyTorch(1.7). It supports many backbones and datasets.

Deployment of 3D-Detection and Tracking pipeline in simulation based on rosbags and real-time.

This project uses PyTorch to classify bone fractures. As well as fine-tuning some famous CNN architectures (like VGG 19, MobileNetV3, RegNet,...), we designed our own architecture. Additionally, we used Transformer architectures (such as Vision Transformer and Swin Transformer). This dataset is Bone Fracture Multi-Region X-ray, available on Kaggle.

An image registration method using convolutional neural network features.

A pytorch implement of RegNet (Designing Netowrk design spaces). Original paper link: https://arxiv.org/pdf/2003.13678.pdf

TF 2.x implementation of RegNet (Designing Network Design Spaces, CVPR 2020).

onnxruntime-web

Optimizing DNN Operators on Mobile GPUs

This project modifies, trains and implements RegNetX-200MF for image classification of a modified variant of ImageNet.

This repository holds the downstream task of Face Mask Classification performed on Self Currated Custom Dataset with various State of the Art deep learning models like ViT, BeIT, DeIT, LeViT, ConvNeXt, VGG16, EfficientNetV2, RegNet and MobileNetV3.

Implementation of RegNet (Tensorflow)

Multi-task perception network for quadruped robots: traversability segmentation, terrain classification and anchor-free object detection in one PyTorch forward pass, exported to ONNX/TensorRT for Jetson Orin.

Creación de una IA capaz de recibir una imagen de entrada y clasificarla según su entorno o contexto (por ejemplo, `Kitchen`, `Office`, etc.). 🔗 Accede a la demo aquí: 👉 https://enviroments-clasification.streamlit.app

Fine-grained art style classification using an attention-enhanced RegNetY-032 with Coordinate Attention and Multi-Head Self-Attention (MHSA) for efficient local-global feature learning.