You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Objective: have an examples directory that is both useful and advertises what MPoL is capable of.
Plan:
structure root README.md with w/ links to completed subdirectories. Relabel subdirectories with 01-, 02-, etc.
create nested snakemake files capable of building all products, with the intention to build locally since significant computational resources are required and otherwise not available using GitHub actions free tier
place appropriate outputs like images in version control and incorporate into .md files to showcase results. If dynamically-updated results are important, investigate using jinja2. But this may just be images, which is fine to do with a hard link.
I also considered resuscitating the Sphinx + Jupyter Notebook + nbsphinx build process, pushing static site to GitHub pages and then assemble + deploy with GitHub actions. However, we initially moved away from this because we felt that most imaging workflows are ill-suited to a Jupyter notebook style workflow, and benefit from being broken up into actual Python scripts. I think the directory approach is more useful, and actually mirrors the way the PyTorch examples are presented.
Publish butterfly example / make more prominent
The text was updated successfully, but these errors were encountered:
Objective: have an examples directory that is both useful and advertises what MPoL is capable of.
Plan:
.md
files to showcase results. If dynamically-updated results are important, investigate using jinja2. But this may just be images, which is fine to do with a hard link.I also considered resuscitating the Sphinx + Jupyter Notebook + nbsphinx build process, pushing static site to GitHub pages and then assemble + deploy with GitHub actions. However, we initially moved away from this because we felt that most imaging workflows are ill-suited to a Jupyter notebook style workflow, and benefit from being broken up into actual Python scripts. I think the directory approach is more useful, and actually mirrors the way the PyTorch examples are presented.
The text was updated successfully, but these errors were encountered: