Skip to content

VMamba: Visual State Space Models,code is based on mamba

Notifications You must be signed in to change notification settings

gabrieldernbach/BrainDeadVMamba

 
 

Repository files navigation

braindead vmamba training

bring your data

train/dog/xxx.png
train/cat/xxy.png
train/fish/[...]/xxz.png

val/dog/123.png
val/cat/nsdf3.png
val/fish/[...]/asd932_.png

test/dog/123.png
test/cat/nsdf3.png
test/fish/[...]/asd932_.png

start training

git clone https://github.com/gabrieldernbach/VMamba
mv my_dataset ./VMamba/dataset
cd VMamba && docker build . -t mamba
docker run -v `pwd`:`pwd` -w `pwd` --gpus all --ipc=host  -it vmamba /bin/bash run.sh
  • requires 24G VRAM, modify run.sh if you happen to have more than 1 gpu
  • training output is saved to results

About

VMamba: Visual State Space Models,code is based on mamba

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 88.5%
  • Cuda 6.9%
  • C++ 3.2%
  • Shell 1.2%
  • Other 0.2%