Skip to main content

MLEM Prototype deployment tool

Project description

# mlem-prototype Project to share code ideas and concepts, track tasks and issues for upcoming MLEM tool

## Examples [DVC Pipeline with mlem](examples/dvc-pipeline/README.md)

## Current state Implemented mlem cli & api

### API #### mlem.api.save Saves object to fs in format of <name>.mlem file and <name> dir with artifacts. .mlem file contains all metadata needed to restore objects and some other fields, like requirements for models or columns and types for data frames #### mlem.api.load Loads object which was saved with mlem.api.save ### CLI #### mlem apply Usage: mlem apply -m <method name> <model> <output> <inputs> Loads model and input data, applies model.method to it and saves result to output path in mlem format.

#### mlem deploy ##### mlem deploy <model> heroku Deploys model to heroku. Needs HEROKU_API_KEY env (get it from heroku.com) and and also this ` REGISTRY_HEROKU_COM_PASSWORD=${HEROKU_API_KEY} REGISTRY_HEROKU_COM_USERNAME=_ ` Deployment metadata is written to .mlem model file (subject to change in future)

##### mlem deploy <model> sagemaker –method predict Deploys model to sagemaker. Need to set aws envs: ` export AWS_ACCESS_KEY_ID= export AWS_SECRET_ACCESS_KEY= export AWS_DEFAULT_REGION=us-east-1 `

##### mlem deploy <model> status Checks status of deployment. For now there is no conventions what it will return

##### mlem deploy <model> destroy Undeploy deployed model. Deployment meta is removed from .mlem file

#### mlem apply-remote Same as mlem apply, but actually sends data to deployed model

#### mlem pack <model> <path> Generate model package to <path>

### API2 #### mlem env create <name> <type> creates new target environment type is one of [sagemaker, heroku]

#### mlem deploy2 <model> <env_name> deploys model to chosen taget env deploy metadata is saved to <model>-<env_name>.deployed.yaml

#### mlem destory2 <deploy-name> destroy deploy described in some <model>-<env_name>.deployed.yaml file

#### mlem status2 <deploy-name> get deployment status

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

mlem-0.1.0.tar.gz (73.7 kB view details)

Uploaded Source

Built Distribution

mlem-0.1.0-py3-none-any.whl (97.8 kB view details)

Uploaded Python 3

File details

Details for the file mlem-0.1.0.tar.gz.

File metadata

  • Download URL: mlem-0.1.0.tar.gz
  • Upload date:
  • Size: 73.7 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/3.4.2 importlib_metadata/4.8.1 pkginfo/1.7.1 requests/2.26.0 requests-toolbelt/0.9.1 tqdm/4.62.3 CPython/3.9.7

File hashes

Hashes for mlem-0.1.0.tar.gz
Algorithm Hash digest
SHA256 16ab2b4ace030636f827b2be6ef871a6caf3da335da713f68337e070d8718bc9
MD5 6c9fed079c396d9e6cf64c0cd40e64f9
BLAKE2b-256 0f57acbca200f0a16a59cb9cafbfb0ad2d65b3848350bb2bd18ba59809347586

See more details on using hashes here.

File details

Details for the file mlem-0.1.0-py3-none-any.whl.

File metadata

  • Download URL: mlem-0.1.0-py3-none-any.whl
  • Upload date:
  • Size: 97.8 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/3.4.2 importlib_metadata/4.8.1 pkginfo/1.7.1 requests/2.26.0 requests-toolbelt/0.9.1 tqdm/4.62.3 CPython/3.9.7

File hashes

Hashes for mlem-0.1.0-py3-none-any.whl
Algorithm Hash digest
SHA256 bdd45a77fc1b0bc9d8f239f9534bf99c6e6dbe62a601b9f2911bb63dd0bc8213
MD5 7d9ab73b34758b2d29a77c304814bd94
BLAKE2b-256 9e9410e12b0d784e48aafb806b94cc82373d9d5cea85bca4ed0b7caec596d7d4

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page