ncmir/cdeep3m repository overview

CDeep3M provides a plug-and-play cloud based deep learning solution for image segmentation of light, electron and X-ray microscopy.

Please see https://github.com/CRBS/cdeep3m⁠ and https://github.com/CRBS/cdeep3m/wiki⁠ for instructions, tutorials and original source code.

The C-11 tags are built on cuda 11 and seem a bit faster

Requirements:

• A system with docker and GPU extensions installed
  • https://docs.docker.com/install/⁠
  • https://docs.docker.com/config/containers/resource_constraints/#gpu⁠
• A Cuda capable GPU with > 11Gb of GPU memory
• System CPU and Memory requirements are very minimal (single core CPU & 2-5Gb RAM)
• Cuda 9 or 10 capable GPU for releases before 2021-12-20 or
• Cuda 10 or 11 capable GPU for releases after 2021-12-20

To quickly get going:

docker run -it --network=host  --gpus all --entrypoint /bin/bash ncmir/cdeep3m
sh test/rundemo.sh