How to reproduce

1. install micromamba

For Linux, macOS, or Git Bash on Windows, install with: "${SHELL}" <(curl -L micro.mamba.pm/install.sh)

Default settings during the installation was fine for me.

2. create environment

micromamba create -n ITCD

if you use HPC, start an interactive session: srun --ntasks=1 --cpus-per-task=1 --mem-per-cpu=2000MB --gpus=1 --partition=short --pty bash -i

micromamba install -n ITCD -c conda-forge r-keras r-terra r-rsample tensorflow-gpu=2.15 "keras<3" cuda=12.2 cudnn=8.9 r-devtools r-sf r-lidr r-ggrepel r-raster=3.6_26 r-patchwork

3. install NeonTreeEvaluation in R

micromamba activate ITCD

R

devtools::install_github("Weecology/NeonTreeEvaluation_package", ref="c4d9953")

3 # do not update any package

quit()

n

micromamba deactivate

4. run the scripts in the created environment

follow instructions in process.sh

5 remove an environemnt if not needed anymore

micromamba remove --name ITCD --all

Licence

This repository is licensed under the GNU General Public License, Version 3, or any later version.

In function.R, the function unet contained modification and the function conv2d_block contained copy from r-tensorflow/unet under the MIT License.

In function.R, the function unet contained modification from keras-team/keras under the Apache License, Version 2.0.