Add pretrained VGGish inference pipeline #6481

	name: Integration Test

	on:
	pull_request:
	branches: [ main ]

	workflow_dispatch:

	jobs:
	build:

	runs-on: ubuntu-18.04
	strategy:
	fail-fast: false
	matrix:
	python-version: [ 3.8 ]

	steps:
	- uses: actions/checkout@v2
	- name: Set up Python ${{ matrix.python-version }}
	uses: actions/setup-python@v2
	with:
	python-version: ${{ matrix.python-version }}
	- name: Install dependencies
	run: \|
	sudo add-apt-repository -y ppa:jonathonf/ffmpeg-4
	sudo apt install -y -qq libavfilter-dev libavdevice-dev
	- name: Install packages
	run: \|
	python -m pip install --quiet --upgrade pip
	python -m pip install --quiet --pre torch -f https://download.pytorch.org/whl/nightly/cpu/torch_nightly.html
	python -m pip install --quiet pytest requests cmake ninja deep-phonemizer sentencepiece flashlight-text git+https://github.com/kpu/kenlm
	python setup.py install
	- name: Run integration test
	run: \|
	cd test && pytest integration_tests -v --use-tmp-hub-dir

Provide feedback