This is a quick example in which you will:

• Train Image classification model based on FashionMNIST dataset using PyTorch framework in Jupyter
• Store the resulting model in MLFlow model registry in ONNX format
• Deploy the model from MLFlow to Triton server

Mentioned above solutions are glued and deployed using Neu.ro MLOps platform.

1. Clone guide, install Neu.ro CLI:

   git clone <this repo> && cd <cloned repo directory>
   pipx install neuro-all
   neuro login

2. Upload the guide notebook and start MLFlow, Triton and Jupyter servers.

   neuro-flow upload ALL
   neuro-flow run mlflow_server
   neuro-flow run triton_server
   neuro-flow run jupyter

3. Follow the quickstart_tutorial.ipynb guide in the oppened Jupyter instance.

   Note: this guide is based on PyTorch Quickstart guide.

4. Cleanup:

   neuro-flow kill ALL
   neuro-flow clean ALL

See Help.md for the detailed Neuro Project Template Reference.