fuse fp8 quant in kv copying and add flashinfer decode mla operator in the attention module #654

	name: Pre-commit checks

	on:
	pull_request:
	branches:
	- main

	jobs:
	pre-commit:
	runs-on: ubuntu-latest

	steps:
	- name: Checkout code
	uses: actions/checkout@v2
	with:
	fetch-depth: 0 # Fetch all history for all branches and tags

	- name: Set up Python
	uses: actions/setup-python@v2
	with:
	python-version: '3.9'

	- name: Install dependencies
	run: \|
	python -m pip install --upgrade pip
	pip install pre-commit
	pre-commit install-hooks

	- name: Run pre-commit on modified files
	run: \|
	if [ -n "$(git diff --name-only ${{ github.event.pull_request.base.sha }} ${{ github.sha }})" ]; then
	pre-commit run --files $(git diff --name-only ${{ github.event.pull_request.base.sha }} ${{ github.sha }})
	else
	echo "No files to check"
	fi

Provide feedback