A main repo for all of our reusable github workflows
When you add a new training script you should add workflows in .github/workflows and in the jobs section it should look like
jobs:
deploy-staging:
uses: viam-modules/common-workflows/.github/workflows/deploy_training_script.yaml@main
with:
framework: tflite
script_name: classification-tflite
model_type: tflite
secrets: inherit
This will call our shared deploy_training_script workflow with the necessary inputs.
When you add a new training script you should add workflows in .github/workflows and in the jobs section it should look like
jobs:
build:
uses: viam-modules/common-workflows/.github/workflows/lint_and_test.yaml@main
with:
test_script_name: scripts/test.sh
container_image: tf:2.16
Inputs:
test_script_name(optional, default:scripts/test.sh): Path to the test script to runcontainer_image(required): The ML training container you will use this script with. Must be one of the supported container names found by calling ListSupportedContainers. An easy way to do this is to runviam train containers list.
These workflows require a few secrets. They have already been set up as github secrets for the viam-modules org. So you shouldn't need to do anything special in your new repo. But for transparency, they can be found:
- API_KEY_ID and API_KEY: This is done using an API key that belongs to the
viam-devenvironment. These can be found by going to app.viam.com -> go to theviam-devorg -> Settings -> Find the keys named GITHUB_VIAM_MODULES. - ORG_ID: This is
organization_idforviam-dev. This can be found by navigating to Settings in app.viam.com and finding theorganization_id.