Skip to content

[Feature] Fix DType casting lazy init #2260

[Feature] Fix DType casting lazy init

[Feature] Fix DType casting lazy init #2260

Workflow file for this run

name: Vmas Tests on Linux
on:
pull_request:
push:
branches:
- nightly
- main
- release/*
workflow_dispatch:
concurrency:
# Documentation suggests ${{ github.head_ref }}, but that's only available on pull_request/pull_request_target triggers, so using ${{ github.ref }}.
# On master, we want all builds to complete even if merging happens faster to make it easier to discover at which point something broke.
group: ${{ github.workflow }}-${{ github.ref == 'refs/heads/main' && format('ci-master-{0}', github.sha) || format('ci-{0}', github.ref) }}
cancel-in-progress: true
jobs:
unittests:
uses: pytorch/test-infra/.github/workflows/linux_job.yml@main
with:
repository: pytorch/rl
runner: "linux.g5.4xlarge.nvidia.gpu"
gpu-arch-type: cuda
gpu-arch-version: "11.7"
timeout: 120
script: |
set -euo pipefail
export PYTHON_VERSION="3.9"
export CU_VERSION="11.7"
export TAR_OPTIONS="--no-same-owner"
export UPLOAD_CHANNEL="nightly"
export TF_CPP_MIN_LOG_LEVEL=0
nvidia-smi
bash .github/unittest/linux_libs/scripts_vmas/setup_env.sh
bash .github/unittest/linux_libs/scripts_vmas/install.sh
bash .github/unittest/linux_libs/scripts_vmas/run_test.sh
bash .github/unittest/linux_libs/scripts_vmas/post_process.sh