Conda environment configurations for deepali libraries.
If not done before, install either Miniconda or Anaconda.
A conda environment for this project can be created using the env Makefile target, i.e.,
make env [NAME=deepali] [EDITABLE=1]
This creates a conda environment with the specified NAME, and subsequently installs the deepali libraries. When EDITABLE=1|true, an editable installation of the libraries is performed using pip install --editable.
Alternatively, run the following conda commands intead of using make:
NAME=deepali
PLATFORM=linux-64
conda create --name $NAME --file environment.$PLATFORM.lock
conda run --name $NAME pip install [--editable] ..
where PLATFORM is one of the following values.
| Platform | Description |
|---|---|
linux-64 |
64-bit Linux distribution (e.g., CentOS 7). |
osx-64 |
64-bit macOS without CUDA device. |
win-64 |
64-bit Microsoft Windows with CUDA device. |
The following tools are required to update the generated dependency files.
- Install Miniconda: Package dependency management tool.
- Install mamba: Faster dependency resolution and more informative error messages.
conda install mamba=0.23 --name base --channel conda-forge
- Install conda-devenv: Advanced conda environment configuration such as conditional dependencies.
conda install conda-devenv=2.3 --name base --channel conda-forge
- Install conda-lock: Lock versions of dependencies and generate explicit lockfiles.
conda install conda-lock=1.0 --name base --channel conda-forge
After editing the environment.devenv.yml file to add, remove, or update required and optional dependencies, run the following command to re-generate the environment.conda-lock.yml and environment.*.lock files.
make all