purbayankar/hybridsn-pytorch

A pytorch based implementation of Hybrid-Spectral-Net as in IEEE GRSL paper "HybridSN: Exploring 3D-2D CNN Feature Hierarchy for Hyperspectral Image Classification".

Hybrid-Spectral-Net for Hyperspectral Image Classification.

Description

The HybridSN is spectral-spatial 3D-CNN followed by spatial 2D-CNN. The 3D-CNN facilitates the joint spatial-spectral feature representation from a stack of spectral bands. The 2D-CNN on top of the 3D-CNN further learns more abstract level spatial representation. Paper link: https://ieeexplore.ieee.org/document/8736016 . Original Keras implimentation: https://github.com/gokriznastic/HybridSN

Model

Fig: HybridSpectralNet (HybridSN) Model with 3D and 2D convolutions for hyperspectral image (HSI) classification.

Requirements

To install requirements:

conda env create -f environment.yml

To download the dataset and setup the folders, run:

bash setup_script.sh

Training

To train the model(s) in the paper, run this command in the A2S2KResNet folder:

python A2S2KResNet.py -d <IN|UP|KSC> -e 200 -i 3 -p 3 -vs 0.9 -o adam

hybridsn-pytorch's Issues

run in pycharm and got this...

RuntimeError: mat1 and mat2 shapes cannot be multiplied (2x64 and 18496x256)

How to solve this problem

RuntimeError: size mismatch, m1: [2 x 64], m2: [18496 x 256] at C:/w/1/s/tmp_conda_3.7_055457/conda/conda-bld/pytorch_1565416617654/work/aten/src\THC/generic/THCTensorMathBlas.cu:273

There were some problems in my preparation.May I ask how to solve the problem?