Test GPU sorter images in singularity on AWS #9

	name: Test GPU sorter images in singularity on AWS

	on:
	workflow_dispatch:

	jobs:
	deploy-runner:
	runs-on: ubuntu-latest
	steps:
	- uses: iterative/setup-cml@v1
	- uses: actions/checkout@v4
	- name: Deploy runner on EC2
	env:
	REPO_TOKEN: ${{ secrets.PERSONAL_ACCESS_TOKEN }}
	AWS_ACCESS_KEY_ID: ${{ secrets.AWS_ACCESS_KEY_ID }}
	AWS_SECRET_ACCESS_KEY: ${{ secrets.AWS_SECRET_ACCESS_KEY }}
	run: \|
	cml runner \
	--cloud=aws \
	--cloud-region=us-east-2 \
	--cloud-type=g4dn.xlarge \
	--labels=cml-gpu
	test-gpu-sorters:
	needs: deploy-runner
	runs-on: [self-hosted, cml-gpu]
	timeout-minutes: 360 # 6h
	container:
	image: docker://iterativeai/cml:0-dvc2-base1-gpu
	options: --privileged --gpus all
	steps:
	- uses: actions/checkout@v4

	- name: Install dependencies
	run: \|
	sudo apt-get update
	sudo apt-get install -y wget squashfs-tools
	python -m pip install -U pip # Official recommended way
	pip install pytest
	pip install -e .[full]
	pip install spython

	- uses: eWaterCycle/setup-singularity@v7
	with:
	singularity-version: 3.8.3

	- name: Run test singularity containers with GPU
	env:
	REPO_TOKEN: ${{ secrets.PERSONAL_ACCESS_TOKEN }}
	run: \|
	pytest -vv --capture=tee-sys -rA src/spikeinterface/sorters/external/tests/test_singularity_containers_gpu.py

Provide feedback