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
mlf-core projects parameters -> shows all possible parameters of a given project nicely formatted.
The parameters are in the MLproject file.
mlf-core projects run -> calls mlflow run, but adds option -A t if docker + allows to directly run Github projects without https://github.com prefix
This is a little bit trickier to implement.
I will just ramble a little bit for now:
MLflow by default expects git related URLs to be explicit -> mlflow run https://github.com/name/repository -> I would like to be able to run mlf-core run name/repository
It would be awesome to have a switch like Nextflow's with-conda or with-docker. Currently users always have to edit the MLproject file themselves to edit in and edit out what they want and don't want. That's annoying. They don't seem to prioritize that over at MLflow.
The -A flags should only be supplied if run with Docker
We should make sure that any additional -A and -P supplied parameters are nicely appended to the run command.
The text was updated successfully, but these errors were encountered:
Zethson
changed the title
Add a command to show all available CLI options for a given project
Add custom mlf-core project command to 1. command to show all available CLI options for a given project 2. replace mlflow run
Jan 28, 2021
mlf-core projects parameters
-> shows all possible parameters of a given project nicely formatted.The parameters are in the MLproject file.
mlf-core projects run
-> calls mlflow run, but adds option -A t if docker + allows to directly run Github projects without https://github.com prefixThis is a little bit trickier to implement.
I will just ramble a little bit for now:
with-conda
orwith-docker
. Currently users always have to edit the MLproject file themselves to edit in and edit out what they want and don't want. That's annoying. They don't seem to prioritize that over at MLflow.The text was updated successfully, but these errors were encountered: