Semantic segmentation models with pretrained convolutional and transformer-based backbones. PyTorch.

PyTorch and TensorFlow/Keras image models with automatic weight conversions and equal API/implementations - Vision Transformer (ViT), ResNetV2, EfficientNetV2, NeRF, SegFormer, MixTransformer, (planned...) DeepLabV3+, ConvNeXtV2, YOLO, etc.

This repository demonstrates how to use TensorFlow based SegFormer model in 🤗 transformers package.

This repository provides an overview of Segformer, architecture encoder in particular. Some details of Segformer can be misleaded, thus makes a short description here to help understand the model. Meanwhile, the code (Keras/TensorFlow) is also provided for supporting.

A Tensorflow Implementation of "SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers"

PyTorch Implementation of SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers

Real time Semantic Segmentation on Cityscapes dataset.

Clothes semantic segmentation with Transformers

Tackling the challenges of off-road environment navigation through Attention Guided Off-Road Semantic Segmentation.

This is a warehouse for SegFormer-pytorch-model, can be used to train your image datasets for segmentation tasks.

Automatic Damage Segmentation of Sanskrit Palm-leaf Manuscripts

Comparing CNN- and Transformer-Based Deep Learning Models for Semantic Segmentation of Remote Sensing Images. Specifically looking at different configurations of U-Net and SegFormer applied to Potsdam, Vaihingen and FloodNet dataset.

On Cloud N: Cloud Cover Detection Challenge Solution

Repository of the master's thesis focused on the Automatic Detection of Objects in the Track

Image Segmentation using Segformer finetuning

A repository of SegFormer model test

This repository contains my project on medical image segmentation using the UW-Madison GI Tract dataset. It will use a Segformer, and include a Gradio app for user-friendly image inference. The model will be deployed on HuggingFace Spaces for easy accessibility, encouraging contributions to advance medical image segmentation in healthcare domain.

Implementation of transformer based SegFormer-B0 from scratch for semantic segmentation.

Pattern Recognition Systems course (4th year, 1st semester)