The Meeshkan Client for interactive machine learning
Project description
Meeshkan client
Getting started
Sign up at app.meeshkan.com and get your token. Add the folder .meeshkan
in your home directory and, inside the folder, add a file named credentials
with the following format:
[meeshkan]
token=my-token
Installation
pip install meeshkan
Command-line interface
To list available commands, execute meeshkan
or meeshkan help
:
Usage: meeshkan [OPTIONS] COMMAND [ARGS]...
Options:
-h, --help Show this message and exit.
--version Show the version and exit.
Commands:
clear Clears Meeshkan log and job directories in ~/.meeshkan.
help Show this message and exit.
list Lists the job queue and status for each job.
sorry Send error logs to Meeshkan HQ.
start Starts Meeshkan service daemon.
status Checks and returns the service daemon status.
stop Stops the service daemon.
submit Submits a new job to the service daemon.
Start service daemon
meeshkan start
If you get Unauthorized
error, please check your credentials. If the problem persists, please contact Meeshkan support.
Check service status
meeshkan status
You should get the message Service is up and running
.
Submit a Python script for execution
Submit the example script hello_world.py for execution:
meeshkan submit [--name job_name] examples/train.py
List submitted jobs
meeshkan list
Stop service
meeshkan stop
Reporting scalars from PyTorch
Get started
Start the service first, then run
meeshkan submit --poll 10 examples/report.py
PyTorch
See examples/pytorch_mnist.py for an example script. To run the script, first ensure that PyTorch is installed in your Python environment. Then submit the example as
meeshkan submit --poll 10 examples/pytorch_mnist.py
Meeshkan reports the training and test loss values to you every 10 seconds.
Development
Installation
For users:
python setup.py install
For developers:
pip install -e .[dev]
Running tests
pytest
OR
python setup.py test
Running lint
pylint -f msvs meeshkan
Building the documentation
cd docs
sphinx-apidoc -f -e -o source/ ../meeshkan/
make html
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for meeshkan-0.1.0-py2.py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | c2e17b1363f01b12946bdc1afbcd1017ec3250d405a0b53e36b84521cc6dfaa0 |
|
MD5 | 824e1d5bd1f7ae2ccbdb248988f03334 |
|
BLAKE2b-256 | de0df30df38d8bcd60f2854312b9adc2f6e2d5de8f5aa0638c033474e5af8068 |