User and Organization Commands

login

Login to cnvrg system

$ cnvrg login

The command will ask for your email and password, and will set default owner

logout

Logout from cnvrg system

$ cnvrg logout

me

shows the current user

$ cnvrg me

Project Commands

new

Create new empty cnvrg project.

$ cnvrg new project_title

The command will create a directory named project_title, with a suggested directories for data science projects.

You will receive a generated link to the web view of the project

options:

--clean, -c : don't create suggested directories, default: false
--docker_image, -d : specify one of cnvrg docker images for this project (for local purposes)

link

Link cnvrg to an existing project (run from inside the project directory)

$ cnvrg link 

The command will link the project to cnvrg, and automatically sync its content.

You will receive a generated link to the web view of the project.

options:

--sync, -s : don't create suggested directories, default:true
--docker_image, -d : specify one of cnvrg docker images for this project (for local purposes)

clone

Clone an existing cnvrg project

$ cnvrg clone URL 

The command will create a directory with the project's name and will download the project's content.

Note: You must have permissions to the project in order to clone it.

status

Shows the current status of the project's tree (similar to git status)

$ cnvrg status 

upload

Uploads current project commit

$ cnvrg upload 

The command will upload the current snapshot of the project as a commit (same as git push).

options:

--ignore, -i : list of files to ignore, can also be specified in .cnvrgignore file
--new_branch, -nb : create new branch
--verbose, -v : print more details when uploading

download

Download latest project's commit

$ cnvrg download

The command will download the latest snapshot of the project (same as git pull).

options:

--verbose, -v : print more details when downloading

sync

Makes download request and then upload request

$ cnvrg sync

options:

--verbose, -v : print more details
--new_branch, -nb : create new branch (for upload)

jump

Jump to a different commit snapshot

$ cnvrg jump 68b793374

The command will first sync the current project status and then jump to the requested commit

Experiments & Notebooks Commands

run

Runs a new experiment

$ cnvrg run python run.py

The command will run the python script on a remote machine,  chosen by the default setting for the organization/user.
A link will be immediately generated for tracking experiment's status and result

options:

--local, -l : run experiment on local machine, default:false
--small, -sm : run experiment on a cpu machine of type small
--medium, -md : run experiment on a cpu machine of type medium
--large, -lg  : run experiment on a cpu machine of type large
--gpu  : run experiment on a gpu machine of type small
--gpuxl : run experiment on a gpu machine of type medium
--gpuxxl : run experiment on a gpu machine of type large
--email_notification, -en : send an email notification when experiment is finished
--commit, -c : run experiment in a specific commit
--schedule, -s : schedule an experiment to run as a Cron job for recurring experiments or on another time, i.e. tomorrow at 7 pm, every Tuesday and Friday at 12
--grid, -g : specify grid yaml for experiment

notebook

Starts a new notebook session

$ cnvrg notebook

The command will start a new notebook session on a remote machine, chosen by the default setting for the organization/user.

A link will be immediately generated for accessing notebook online

options:

--local, -l : run experiment on local machine, default:false
--small, -sm : run experiment on a cpu machine of type small
--medium, -md : run experiment on a cpu machine of type medium
--large, -lg  : run experiment on a cpu machine of type large
--gpu  : run experiment on a gpu machine of type small
--gpuxl : run experiment on a gpu machine of type medium
--gpuxxl : run experiment on a gpu machine of type large

Datasets Commands

data init

Creates a new dataset from an existing (local) directory

$ cnvrg data init

You will receive a link to your new created dataset.

data upload

Uploads the contents of the directory

$ cnvrg data upload

data clone

Clone a cnvrg dataset locally

$ cnvrg data clone https://dataset_url

Did this answer your question?