Fuego
Project description
fuego
A 🔥 tool for running code in the cloud
Note
❗ This project is just an idea right now and is being actively developed ❗
The idea
A unified interface to view/manage runs, compute instances, data, etc. across your authenticated cloud providers.
Currently, this project only works for AzureML training runs. The idea is to support other cloud providers (AWS, GCP, etc.) and other run types (inference, data preparation, etc.) in the future.
Installation
For now, you can install from source:
git clone https://github.com/huggingface/fuego.git
cd fuego
pip install -e ".[azureml]"
WIP API
The Python API and CLI should have very similar experiences so folks can use whichever they prefer.
Python
import fuego
fuego.run(
# Fuego Run Args - these are the same across providers
script='./examples/simple_example_with_requirements/run.py',
provider='azureml',
instance_type='cpu',
instance_count=1,
requirements_file='./examples/simple_example_with_requirements/requirements.txt',
# Script kwargs - these are passed to the script as argparse args
message='Howdy, world!',
)
CLI
fuego run \
--provider azureml \
--instance-type cpu \
--instance-count 1 \
--requirements-file ./examples/simple_example_with_requirements/requirements.txt \
./examples/simple_example_with_requirements/run.py \
--message "'Howdy, world!'"
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
fuego-0.0.6rc0.tar.gz
(10.4 kB
view hashes)
Built Distribution
fuego-0.0.6rc0-py3-none-any.whl
(10.8 kB
view hashes)