docker-test-gpu #13

	name: docker-test-gpu

	on:
	workflow_dispatch:
	inputs:
	backend:
	type: choice
	description: Backend type
	options:
	- llama-cuda
	- exllama
	- exllama2-gptq
	- exllama2-exl2

	permissions: read-all

	jobs:
	test:
	runs-on: self-hosted
	timeout-minutes: 240
	steps:
	- uses: AutoModality/action-clean@d004b47eb728f2a83316fc70ba9a62ef97278013 # v1.1.0
	- uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11 # v4.1.1

	- name: build aikit
	run: \|
	docker buildx build . -t aikit:test \
	--load --progress plain

	- name: build test model
	run: \|
	docker buildx build . -t testmodel:test \
	-f test/aikitfile-${{ github.event.inputs.backend }}.yaml \
	--load --progress plain

	- name: list images
	run: docker images

	- name: run test model
	run: docker run --name testmodel -d --rm -p 8080:8080 --gpus all testmodel:test

	- name: install e2e dependencies
	run: make test-e2e-dependencies

	- name: run test
	run: \|
	curl http://127.0.0.1:8080/v1/chat/completions -H "Content-Type: application/json" -d '{
	"model": "llama-2-7b-chat",
	"messages": [{"role": "user", "content": "explain kubernetes in a sentence"}]
	}'

	- run: docker stop testmodel

	- name: save logs
	if: always()
	run: docker logs testmodel > /tmp/docker-${{ github.event.inputs.backend }}.log

	- name: publish test artifacts
	if: always()
	uses: actions/upload-artifact@c7d193f32edcb7bfad88892161225aeda64e9392 # v4.0.0
	with:
	name: test-${{ github.event.inputs.backend }}
	path: \|
	/tmp/*.log

Provide feedback