Explore projects

Unified framework for comparing Vision Transformers for Image Segmentation

Unified framework for comparing Vision Transformers for Image Segmentation

Encoder decoder architecture for semantic segmentation .Encoder consists of a pretrained resnet18 and the decoder is composed of multiple layers of transposed convolution.

This is mmdetection library that implements deformable convolution neural networks. With this library I can use attention layers with mmdetection implemented in pytorch