Skip to content
Back to pull request #1909

[BugFix] Make sure ParallelEnv does not overflow mem when policy requires grad #614

Sign in to view logs

[BugFix] Make sure ParallelEnv does not overflow mem when policy requires grad

[BugFix] Make sure ParallelEnv does not overflow mem when policy requires grad #614

Workflow file for this run

.github/workflows/test-linux-libs.yml at a6dc5ee
name: Libs 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-atari-dqn:
strategy:
matrix:
python_version: ["3.9"]
cuda_arch_version: ["12.1"]
if: ${{ github.event_name == 'push' || contains(github.event.pull_request.labels.*.name, 'Data') }}
uses: pytorch/test-infra/.github/workflows/linux_job.yml@main
with:
repository: pytorch/rl
runner: "linux.g5.4xlarge.nvidia.gpu"
docker-image: "nvidia/cudagl:11.4.0-base"
timeout: 120
script: |
set -euo pipefail
export PYTHON_VERSION="3.9"
export CU_VERSION="cu117"
export TAR_OPTIONS="--no-same-owner"
export UPLOAD_CHANNEL="nightly"
export TF_CPP_MIN_LOG_LEVEL=0
bash .github/unittest/linux_libs/scripts_ataridqn/setup_env.sh
bash .github/unittest/linux_libs/scripts_ataridqn/install.sh
bash .github/unittest/linux_libs/scripts_ataridqn/run_test.sh
bash .github/unittest/linux_libs/scripts_ataridqn/post_process.sh
unittests-brax:
strategy:
matrix:
python_version: ["3.9"]
cuda_arch_version: ["12.1"]
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="12.1"
export TAR_OPTIONS="--no-same-owner"
export UPLOAD_CHANNEL="nightly"
export TF_CPP_MIN_LOG_LEVEL=0
export BATCHED_PIPE_TIMEOUT=60
nvidia-smi
bash .github/unittest/linux_libs/scripts_brax/run_all.sh
unittests-d4rl:
strategy:
matrix:
python_version: ["3.9"]
cuda_arch_version: ["12.1"]
if: ${{ github.event_name == 'push' || contains(github.event.pull_request.labels.*.name, 'Data') }}
uses: pytorch/test-infra/.github/workflows/linux_job.yml@main
with:
repository: pytorch/rl
runner: "linux.g5.4xlarge.nvidia.gpu"
docker-image: "nvidia/cudagl:11.4.0-base"
timeout: 120
script: |
set -euo pipefail
export PYTHON_VERSION="3.9"
export CU_VERSION="cu117"
export TAR_OPTIONS="--no-same-owner"
export UPLOAD_CHANNEL="nightly"
export TF_CPP_MIN_LOG_LEVEL=0
export BATCHED_PIPE_TIMEOUT=60
bash .github/unittest/linux_libs/scripts_d4rl/setup_env.sh
bash .github/unittest/linux_libs/scripts_d4rl/install.sh
bash .github/unittest/linux_libs/scripts_d4rl/run_test.sh
bash .github/unittest/linux_libs/scripts_d4rl/post_process.sh
unittests-envpool:
strategy:
matrix:
python_version: ["3.9"]
cuda_arch_version: ["12.1"]
if: ${{ github.event_name == 'push' || contains(github.event.pull_request.labels.*.name, 'Data') }}
uses: pytorch/test-infra/.github/workflows/linux_job.yml@main
with:
repository: pytorch/rl
runner: "linux.g5.4xlarge.nvidia.gpu"
docker-image: "nvidia/cudagl:11.4.0-base"
timeout: 120
script: |
set -euo pipefail
export PYTHON_VERSION="3.9"
export CU_VERSION="cu117"
export TAR_OPTIONS="--no-same-owner"
export UPLOAD_CHANNEL="nightly"
export TF_CPP_MIN_LOG_LEVEL=0
export BATCHED_PIPE_TIMEOUT=60
bash .github/unittest/linux_libs/scripts_d4rl/setup_env.sh
bash .github/unittest/linux_libs/scripts_d4rl/install.sh
bash .github/unittest/linux_libs/scripts_d4rl/run_test.sh
bash .github/unittest/linux_libs/scripts_d4rl/post_process.sh
unittests-gendgrl:
strategy:
matrix:
python_version: ["3.9"]
cuda_arch_version: ["12.1"]
if: ${{ github.event_name == 'push' || contains(github.event.pull_request.labels.*.name, 'Data') }}
uses: pytorch/test-infra/.github/workflows/linux_job.yml@main
with:
repository: pytorch/rl
runner: "linux.g5.4xlarge.nvidia.gpu"
docker-image: "nvidia/cudagl:11.4.0-base"
timeout: 120
script: |
set -euo pipefail
export PYTHON_VERSION="3.9"
export CU_VERSION="cu117"
export TAR_OPTIONS="--no-same-owner"
export UPLOAD_CHANNEL="nightly"
export TF_CPP_MIN_LOG_LEVEL=0
bash .github/unittest/linux_libs/scripts_gen-dgrl/setup_env.sh
bash .github/unittest/linux_libs/scripts_gen-dgrl/install.sh
bash .github/unittest/linux_libs/scripts_gen-dgrl/run_test.sh
bash .github/unittest/linux_libs/scripts_gen-dgrl/post_process.sh
unittests-gym:
strategy:
matrix:
python_version: ["3.9"]
cuda_arch_version: ["12.1"]
uses: pytorch/test-infra/.github/workflows/linux_job.yml@main
with:
repository: pytorch/rl
runner: "linux.g5.4xlarge.nvidia.gpu"
# gpu-arch-type: "cuda"
# gpu-arch-version: "11.6"
docker-image: "nvidia/cudagl:11.4.0-base"
timeout: 120
script: |
set -euxo pipefail
export PYTHON_VERSION="3.9"
# export CU_VERSION="${{ inputs.gpu-arch-version }}"
export CU_VERSION="11.4"
export LD_LIBRARY_PATH="$LD_LIBRARY_PATH:/work/mujoco-py/mujoco_py/binaries/linux/mujoco210/bin"
export TAR_OPTIONS="--no-same-owner"
export BATCHED_PIPE_TIMEOUT=60
./.github/unittest/linux_libs/scripts_gym/setup_env.sh
./.github/unittest/linux_libs/scripts_gym/batch_scripts.sh
./.github/unittest/linux_libs/scripts_gym/post_process.sh
unittests-jumanji:
strategy:
matrix:
python_version: ["3.9"]
cuda_arch_version: ["12.1"]
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="12.1"
export TAR_OPTIONS="--no-same-owner"
export UPLOAD_CHANNEL="nightly"
export TF_CPP_MIN_LOG_LEVEL=0
export BATCHED_PIPE_TIMEOUT=60
nvidia-smi
bash .github/unittest/linux_libs/scripts_jumanji/setup_env.sh
bash .github/unittest/linux_libs/scripts_jumanji/install.sh
bash .github/unittest/linux_libs/scripts_jumanji/run_test.sh
bash .github/unittest/linux_libs/scripts_jumanji/post_process.sh
unittests-minari:
strategy:
matrix:
python_version: ["3.9"]
cuda_arch_version: ["12.1"]
if: ${{ github.event_name == 'push' || contains(github.event.pull_request.labels.*.name, 'Data') }}
uses: pytorch/test-infra/.github/workflows/linux_job.yml@main
with:
repository: pytorch/rl
runner: "linux.g5.4xlarge.nvidia.gpu"
docker-image: "nvidia/cudagl:11.4.0-base"
timeout: 120
script: |
set -euo pipefail
export PYTHON_VERSION="3.9"
export CU_VERSION="cu117"
export TAR_OPTIONS="--no-same-owner"
export UPLOAD_CHANNEL="nightly"
export TF_CPP_MIN_LOG_LEVEL=0
export BATCHED_PIPE_TIMEOUT=60
bash .github/unittest/linux_libs/scripts_minari/setup_env.sh
bash .github/unittest/linux_libs/scripts_minari/install.sh
bash .github/unittest/linux_libs/scripts_minari/run_test.sh
bash .github/unittest/linux_libs/scripts_minari/post_process.sh
unittests-openx:
strategy:
matrix:
python_version: ["3.9"]
cuda_arch_version: ["12.1"]
if: ${{ github.event_name == 'push' || contains(github.event.pull_request.labels.*.name, 'Data') }}
uses: pytorch/test-infra/.github/workflows/linux_job.yml@main
with:
repository: pytorch/rl
runner: "linux.g5.4xlarge.nvidia.gpu"
docker-image: "nvidia/cudagl:11.4.0-base"
timeout: 120
script: |
set -euo pipefail
export PYTHON_VERSION="3.9"
export CU_VERSION="cu117"
export TAR_OPTIONS="--no-same-owner"
export UPLOAD_CHANNEL="nightly"
export TF_CPP_MIN_LOG_LEVEL=0
export BATCHED_PIPE_TIMEOUT=60
bash .github/unittest/linux_libs/scripts_openx/setup_env.sh
bash .github/unittest/linux_libs/scripts_openx/install.sh
bash .github/unittest/linux_libs/scripts_openx/run_test.sh
bash .github/unittest/linux_libs/scripts_openx/post_process.sh
unittests-pettingzoo:
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="12.1"
export TAR_OPTIONS="--no-same-owner"
export UPLOAD_CHANNEL="nightly"
export TF_CPP_MIN_LOG_LEVEL=0
export BATCHED_PIPE_TIMEOUT=60
nvidia-smi
bash .github/unittest/linux_libs/scripts_pettingzoo/setup_env.sh
bash .github/unittest/linux_libs/scripts_pettingzoo/install.sh
bash .github/unittest/linux_libs/scripts_pettingzoo/run_test.sh
bash .github/unittest/linux_libs/scripts_pettingzoo/post_process.sh
unittests-robohive:
strategy:
matrix:
python_version: ["3.9"]
cuda_arch_version: ["12.1"]
uses: pytorch/test-infra/.github/workflows/linux_job.yml@main
with:
repository: pytorch/rl
runner: "linux.g5.4xlarge.nvidia.gpu"
docker-image: "nvidia/cudagl:11.4.0-base"
timeout: 120
script: |
set -euo pipefail
export PYTHON_VERSION="3.9"
export CU_VERSION="cu117"
export TAR_OPTIONS="--no-same-owner"
export UPLOAD_CHANNEL="nightly"
export TF_CPP_MIN_LOG_LEVEL=0
export BATCHED_PIPE_TIMEOUT=60
bash .github/unittest/linux_libs/scripts_robohive/setup_env.sh
bash .github/unittest/linux_libs/scripts_robohive/install_and_run_test.sh
bash .github/unittest/linux_libs/scripts_robohive/post_process.sh
unittests-roboset:
strategy:
matrix:
python_version: ["3.9"]
cuda_arch_version: ["12.1"]
if: ${{ github.event_name == 'push' || contains(github.event.pull_request.labels.*.name, 'Data') }}
uses: pytorch/test-infra/.github/workflows/linux_job.yml@main
with:
repository: pytorch/rl
runner: "linux.g5.4xlarge.nvidia.gpu"
docker-image: "nvidia/cudagl:11.4.0-base"
timeout: 120
script: |
set -euo pipefail
export PYTHON_VERSION="3.9"
export CU_VERSION="cu117"
export TAR_OPTIONS="--no-same-owner"
export UPLOAD_CHANNEL="nightly"
export TF_CPP_MIN_LOG_LEVEL=0
export BATCHED_PIPE_TIMEOUT=60
bash .github/unittest/linux_libs/scripts_roboset/setup_env.sh
bash .github/unittest/linux_libs/scripts_roboset/install.sh
bash .github/unittest/linux_libs/scripts_roboset/run_test.sh
bash .github/unittest/linux_libs/scripts_roboset/post_process.sh
unittests-sklearn:
strategy:
matrix:
python_version: ["3.9"]
cuda_arch_version: ["12.1"]
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"
docker-image: "nvidia/cudagl:11.4.0-base"
timeout: 120
script: |
set -euo pipefail
export PYTHON_VERSION="3.9"
export CU_VERSION="cu117"
export TAR_OPTIONS="--no-same-owner"
export UPLOAD_CHANNEL="nightly"
export TF_CPP_MIN_LOG_LEVEL=0
export BATCHED_PIPE_TIMEOUT=60
bash .github/unittest/linux_libs/scripts_sklearn/setup_env.sh
bash .github/unittest/linux_libs/scripts_sklearn/install.sh
bash .github/unittest/linux_libs/scripts_sklearn/run_test.sh
bash .github/unittest/linux_libs/scripts_sklearn/post_process.sh
unittests-smacv2:
strategy:
matrix:
python_version: ["3.9"]
cuda_arch_version: ["12.1"]
if: ${{ github.event_name == 'push' || contains(github.event.pull_request.labels.*.name, 'Environments') }}
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="12.1"
export TAR_OPTIONS="--no-same-owner"
export UPLOAD_CHANNEL="nightly"
export TF_CPP_MIN_LOG_LEVEL=0
export BATCHED_PIPE_TIMEOUT=60
nvidia-smi
bash .github/unittest/linux_libs/scripts_smacv2/setup_env.sh
bash .github/unittest/linux_libs/scripts_smacv2/install.sh
bash .github/unittest/linux_libs/scripts_smacv2/run_test.sh
bash .github/unittest/linux_libs/scripts_smacv2/post_process.sh
unittests-vd4rl:
strategy:
matrix:
python_version: ["3.9"]
cuda_arch_version: ["12.1"]
if: ${{ github.event_name == 'push' || contains(github.event.pull_request.labels.*.name, 'Data') }}
uses: pytorch/test-infra/.github/workflows/linux_job.yml@main
with:
repository: pytorch/rl
runner: "linux.g5.4xlarge.nvidia.gpu"
docker-image: "nvidia/cudagl:11.4.0-base"
timeout: 120
script: |
set -euo pipefail
export PYTHON_VERSION="3.9"
export CU_VERSION="cu117"
export TAR_OPTIONS="--no-same-owner"
export UPLOAD_CHANNEL="nightly"
export TF_CPP_MIN_LOG_LEVEL=0
export BATCHED_PIPE_TIMEOUT=60
bash .github/unittest/linux_libs/scripts_vd4rl/setup_env.sh
bash .github/unittest/linux_libs/scripts_vd4rl/install.sh
bash .github/unittest/linux_libs/scripts_vd4rl/run_test.sh
bash .github/unittest/linux_libs/scripts_vd4rl/post_process.sh
unittests-vmas:
strategy:
matrix:
python_version: ["3.9"]
cuda_arch_version: ["12.1"]
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="12.1"
export TAR_OPTIONS="--no-same-owner"
export UPLOAD_CHANNEL="nightly"
export TF_CPP_MIN_LOG_LEVEL=0
export BATCHED_PIPE_TIMEOUT=60
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