Skip to content

Latest commit

 

History

History
15 lines (11 loc) · 551 Bytes

File metadata and controls

15 lines (11 loc) · 551 Bytes

#Clipper-dev

Deployed an image classifier built using pytorch on Clipper

Install: 1. Clipper 2. Pytorch 3. Seaborn

  1. Start Docker and minikube (via minikube start --vm-driver hyperkit)
  2. Init Clipper & deploy model: run clipper_dev.py to deploy pytorch neural network, run clipper_randop.py to deploy a model that returns a random vector
  3. Open batch_client.ipynb in a jupyter notebook. Change the filename to save the plot to a different file.
  4. Run stop_clipper.py to stop Clipper.

Visit www.clipper.ai for more information about Clipper