Skip to main content

modelstore is a library for versioning, exporting, and storing machine learning models

Project description

modelstore

modelstore is a Python library that allows you to version, export, save and download machine learning models in your choice of storage.

For more details, please refer to the documentation.

This library has been developed using Python 3.6, 3.7 and 3.8. Please open an issue if you have any trouble!

Installation

pip install modelstore

Downloads

Supported storage types

  • AWS S3 Bucket (example)
  • Azure Blob Storage (example)
  • Google Cloud Storage Bucket (example)
  • A filesystem directory (example)
  • A hosted storage option (Get in touch for an API key! example)

Supported machine learning libraries

Is there a machine learning framework that is missing? Feel free to open an issue or send us a pull request!

Features

  • Model versioning on each upload
  • Replaces all the boiler plate code you need to save models
  • Support for multiple clouds (AWS, GCP, Azure)
  • List models in a domain
  • Download model archives by id
  • Create model states and manage which state a model is in

Example Usage

Colab Notebook

There is a full example in this Colab notebook.

Python Script

from modelstore import ModelStore
# And your other imports

# Train your model
clf = RandomForestClassifier(n_estimators=10)
clf = clf.fit(X, Y)

# Create a model store that uses a one of the storage options
# In this example, the model store is created with a GCP bucket
model_store = ModelStore.from_gcloud(
   project_name="my-project",
   bucket_name="my-bucket",
)

# Upload the archive to your model store
domain = "example-model"
meta_data = model_store.upload(domain, model=clf)

# Print the meta-data about the model
print(json.dumps(meta_data, indent=4))

# Download the model back!
target = f"downloaded-{model_type}-model"
os.makedirs(target, exist_ok=True)
model_path = model_store.download(
   local_path=target,
   domain=model_domain,
   model_id=meta["model"]["model_id"],
)

Recorded demo from Data Talks Club

This interview and demo, recorded with Alexey from the Data Talks Club in July 2021, is based on modelstore==0.0.6. Click below to watch on YouTube.

Demo on YouTube

License

Copyright 2020 Neal Lathia

Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at

http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.

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

modelstore-0.0.7.tar.gz (42.5 kB view hashes)

Uploaded Source

Built Distribution

modelstore-0.0.7-py3-none-any.whl (85.7 kB view hashes)

Uploaded Python 3

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