Skip to content

Latest commit

 

History

History
17 lines (14 loc) · 438 Bytes

ch4-execute-learning.md

File metadata and controls

17 lines (14 loc) · 438 Bytes

Start Traininig

Push the pretrained model to global-model at master branch.

Push a training plan, named plan.json, to train-plan repository.

{
    "name": "MNIST",
    "round": 100,
    "edge": 2,
    "EpR": 1,
    "timeout": 86400,
    "pretrainedModel": "master"
}

There will be 100 commits in global-model, local-model1 and local-model2 respectively. And a tag will be on the last commit of global-model.