Klever Model Registry

English | 中文

Klever Model Registry is a Cloud Native ML model registry. Use Klever Model Registry in order to:

• Manage your ML models
• Version and deliver your ML models with the existing infrastructures
• Keep track of ML models' hyperparameters and so on to help decision makers
• Convert models between different formats ( e.g. TensorFlow SavedModel, ONNX )
• Serve the model
• Get the standalone executable program to deploy ML model inference services on edge devices/servers ( Coming Soon! )

Klever Model Registry's features:

• Deploy with Docker and Kubernetes
• Keep non-Invasive for your business
• Manage ML models like Docker ( With the help of kleveross/ormb )
• Reuse Harbor to store models, without any new infrastructure
• Support model serving for models managed by MLflow
• Extract models signatures for:
  • SavedModel
  • ONNX
  • GraphDef
  • NetDef
  • Keras H5
  • CaffeModel
  • TorchScript
  • MXNetParams
  • PMML
• Convert models from:
  • MXNetParams to ONNX
  • Keras H5 to SavedModel
  • CaffeModel to NetDef

See our official documentations for more information。

UI MockUp

Installation

Build the image

Clone:

$ git clone https://github.com/kleveross/klever-model-registry
$ cd klever-model-registry

Get the dependencies:

$ go mod tidy

Build:

$ make docker-build

Install

Please have a look at docs/installation.md.

If you want to trial quickly, you can run installation script as follow.

$ wget https://raw.githubusercontent.com/kleveross/klever-model-registry/master/scripts/installation/install-k8s-1.14.sh
$ bash install-k8s-1.14.sh <master-ip>

or

$ wget https://raw.githubusercontent.com/kleveross/klever-model-registry/master/scripts/installation/install-k8s-1.19.sh
$ bash install-k8s-1.19.sh <master-ip>

Community

klever-model-registry project is part of Klever, a Cloud Native Machine Learning platform.

The Klever slack workspace is klever.slack.com. To join, click this invitation to our Slack workspace.