fix the bug of stop training around 4 hours. #68

	name: ci
	on:
	pull_request:
	push:
	branches: [main, feat/Sana-public, feat/Sana-public-for-NVLab]
	concurrency:
	group: ci-${{ github.workflow }}-${{ github.ref }}
	cancel-in-progress: true
	# if: ${{ github.repository == 'Efficient-Large-Model/Sana' }}
	jobs:
	pre-commit:
	runs-on: ubuntu-latest
	steps:
	- name: Check out Git repository
	uses: actions/checkout@v4
	- name: Set up Python
	uses: actions/setup-python@v5
	with:
	python-version: 3.10.10
	- name: Test pre-commit hooks
	uses: pre-commit/[email protected]
	tests-bash:
	# needs: pre-commit
	runs-on: self-hosted
	steps:
	- name: Check out Git repository
	uses: actions/checkout@v4
	- name: Set up Python
	uses: actions/setup-python@v5
	with:
	python-version: 3.10.10
	- name: Set up the environment
	run: \|
	bash environment_setup.sh
	- name: Run tests with Slurm
	run: \|
	sana-run --pty -m ci -J tests-bash bash tests/bash/entry.sh

	# tests-python:
	# needs: pre-commit
	# runs-on: self-hosted
	# steps:
	# - name: Check out Git repository
	# uses: actions/checkout@v4
	# - name: Set up Python
	# uses: actions/setup-python@v5
	# with:
	# python-version: 3.10.10
	# - name: Set up the environment
	# run: \|
	# ./environment_setup.sh
	# - name: Run tests with Slurm
	# run: \|
	# sana-run --pty -m ci -J tests-python pytest tests/python

Provide feedback