environment setup for floyd cloud

Floyd_Cloud Set Up.md

Login - Initialize - Run

Sign Up first

Optionally we can do all things to create a conda environment.

Login

1. local installation using 'pip install -U floyd-cli'
2. floyd login -u <user_name>

Initialize

If create conda environment, now it's time to activate

1. First go to your floyd account and create a new project
2. Find the initialize commnad, that generate after creating new project. We need to run this command later, just remember
3. Navigate to your local project directory which you want to run on cloud, using cd .. command we can go to those directory 3.1. We can also clone git repo here, if our project hosted on github server
4. Now the project directory, initialize the command which we get after creating project floyd. Something like following

• floyd init <user_name>/<project_name>

5. Now, if our project has DataSet, then following things to do
6. First create a data sets project on floyd and by this we get some initialize command.. (same things before)
7. Navigate to local data sets folder and run above found command to initialize
8. We need to upload the datasets now by using following command

• floyd data upload After that data sets should running on the server.

9. Now we need to mount the dat sets. Now, go to project directory on local machine and run the following command

• floyd run --data <user_name>/datasets/<datasets_name_set_while_creating_in_floyd>/1:<datasets_name_set_while_creating_in_floyd> --mode jupyter

Run

Navigate to project directory and run following command

• floyd run -mode jupyter --env tensorflow on CPU
• floyd run -mode jupyter --gpu -env tensorflow on GPU *Note: There are many flag for running project.