Refactor inference optimizations #862

	name: push and pull request testing
	on:
	push:
	branches:
	- main
	pull_request:
	branches:
	- main

	jobs:
	build:
	strategy:
	max-parallel: 8
	fail-fast: false
	matrix:
	python-version: [3.7, 3.8]
	platform: [ubuntu-latest, macos-latest]

	runs-on: ${{ matrix.platform }}

	steps:
	- uses: actions/checkout@v2
	- name: Set up Python ${{ matrix.python-version }}
	uses: actions/setup-python@v1
	with:
	python-version: ${{ matrix.python-version }}
	- name: Install dependencies
	run: \|
	- name: Pip upgrade
	run: python -m pip install --upgrade pip
	- name: Sockeye requirements
	run: pip install -r requirements/requirements.txt
	- name: DeepSpeed requirements
	run: pip install -r requirements/requirements.deepspeed.txt
	- name: Faiss requirements
	run: \|
	if [ "$RUNNER_OS" == "Linux" ]; then
	pip install -r requirements/requirements.faiss-cpu.txt
	fi
	shell: bash
	- name: Development requirements
	run: pip install -r requirements/requirements.dev.txt
	- name: Unit tests
	run: \|
	pytest --version
	pytest
	- name: Pylint
	run: \|
	pylint --version
	pylint --rcfile=.pylintrc sockeye
	pylint --rcfile=.pylintrc test --disable=E1102
	- name: MyPy
	run: \|
	mypy --version
	mypy --ignore-missing-imports --follow-imports=silent @typechecked-files --no-strict-optional
	- name: Check manifest
	run: check-manifest --ignore sockeye/git_version.py
	- name: System tests
	run: \|
	pytest test/system

Provide feedback