jaynes 0.5.0

A tool for running python code with runner on aws

Todo

• [x] get the initial template to work

Done

Installation

pip install jaynes

Usage (Show me the Mo-NAY!! :moneybag::money_with_wings:)

Check out the example folder for projects that you can run.

Low-level Usage (Deprecated)

Take a look at the folder test_projects/! These project scripts are used for test and development, so they should work out-of-the-box (if you have the right box ahem).

To run Local Docker

from jaynes import Jaynes, templates
from main import train, RUN

S3_PREFIX = "s3://ge-bair/code-block/new-experiment/"
J = Jaynes(
    launch_log="jaynes_launch.log",
    mounts=[
        templates.S3Mount(local="https://github.com/episodeyang/jaynes/blob/master/", s3_prefix=S3_PREFIX, pypath=True),
        templates.S3Mount(local=".https://github.com/episodeyang/jaynes/blob/master/.https://github.com/episodeyang/jaynes/blob/master/", s3_prefix=S3_PREFIX, pypath=True, file_mask="https://github.com/episodeyang/jaynes/blob/master/__init__.py https://github.com/episodeyang/jaynes/blob/master/jaynes"),
        templates.S3UploadMount(docker_abs=RUN.log_dir, s3_prefix=S3_PREFIX, local=RUN.log_dir, sync_s3=True)
    ],
)
J.set_runner(
    docker=templates.DockerRun("python:3.6",
                               pypath=":".join([m.pypath for m in J.mounts if hasattr(m, "pypath") and m.pypath]),
                               docker_startup_scripts=("pip install cloudpickle",),
                               docker_mount=" ".join([m.docker_mount for m in J.mounts]),
                               use_gpu=False).run(train, log_dir=RUN.log_dir)
)
J.run_local_setup(verbose=True)
J.launch_local_docker(verbose=True, delay=30)

To run on remote server with ssh

from jaynes import Jaynes, templates
from main import train, RUN

S3_PREFIX = "s3://ge-bair/code-block/run_remote/"

output_mount = templates.S3UploadMount(docker_abs=RUN.log_dir, s3_prefix=S3_PREFIX, local=RUN.log_dir, sync_s3=True)

J = Jaynes(
    launch_log="jaynes_launch.log",
    mounts=[
        templates.S3Mount(local="https://github.com/episodeyang/jaynes/blob/master/", s3_prefix=S3_PREFIX, pypath=True),
        templates.S3Mount(local=".https://github.com/episodeyang/jaynes/blob/master/.https://github.com/episodeyang/jaynes/blob/master/", s3_prefix=S3_PREFIX, pypath=True, file_mask="https://github.com/episodeyang/jaynes/blob/master/__init__.py https://github.com/episodeyang/jaynes/blob/master/jaynes"),
        output_mount
    ],
)
J.set_runner(
    docker=templates.DockerRun("python:3.6",
                               pypath=":".join([m.pypath for m in J.mounts if hasattr(m, "pypath") and m.pypath]),
                               docker_startup_scripts=("pip install cloudpickle",),
                               docker_mount=" ".join([m.docker_mount for m in J.mounts]),
                               use_gpu=False).run(train, log_dir=RUN.log_dir)
)
J.run_local_setup(verbose=True)
J.make_launch_script(log_dir=output_mount.remote_abs, instance_tag=RUN.instance_prefix, sudo=True,
                     terminate_after_finish=True)
J.launch_ssh(ip_address="52.11.206.135", pem="~/.ec2/ge-berkeley.pem", script_dir=output_mount.remote_abs, verbose=True)

To Launch EC2 instance and Run on Startup

make sure the instance terminate after docker run!

from jaynes import Jaynes, templates
from main import train, RUN

S3_PREFIX = "s3://ge-bair/code-block/run_ssh/"

output_mount = templates.S3UploadMount(docker_abs=RUN.log_dir, s3_prefix=S3_PREFIX, local=RUN.log_dir, sync_s3=True)

J = Jaynes(
    launch_log="jaynes_launch.log",
    mounts=[
        templates.S3Mount(local="https://github.com/episodeyang/jaynes/blob/master/", s3_prefix=S3_PREFIX, pypath=True),
        templates.S3Mount(local=".https://github.com/episodeyang/jaynes/blob/master/.https://github.com/episodeyang/jaynes/blob/master/", s3_prefix=S3_PREFIX, pypath=True, file_mask="https://github.com/episodeyang/jaynes/blob/master/__init__.py https://github.com/episodeyang/jaynes/blob/master/jaynes"),
        output_mount
    ],
)
J.set_runner(
    docker=templates.DockerRun("python:3.6",
                               pypath=":".join([m.pypath for m in J.mounts if hasattr(m, "pypath") and m.pypath]),
                               docker_startup_scripts=("pip install cloudpickle",),
                               docker_mount=" ".join([m.docker_mount for m in J.mounts]),
                               use_gpu=False).run(train, log_dir=RUN.log_dir)
)
J.run_local_setup(verbose=True)
J.make_launch_script(log_dir=output_mount.remote_abs, instance_tag=RUN.instance_prefix, sudo=False,
                     terminate_after_finish=True)
J.launch_ec2(region="us-west-2", image_id="ami-bd4fd7c5", instance_type="p2.xlarge", key_name="ge-berkeley",
             security_group="torch-gym-prebuilt", spot_price=None,
             iam_instance_profile_arn="arn:aws:iam::055406702465:instance-profile/main", dry=False)

Jaynes does the following:

To Develop

git clone https://github.com/episodeyang/jaynes.git
cd jaynes
make dev

To test, run

make test

This make dev command should build the wheel and install it in your current python environment. Take a look at the https://github.com/episodeyang/jaynes/blob/master/Makefile for details.

To publish, first update the version number, then do:

make publish

Acknowledgements

This code-block is inspired by @justinfu’s doodad, which is in turn built on top of Peter Chen’s script.

This code-block is written from scratch to allow a more permissible open-source license (BSD). Go bears :bear: !!