Make test-maxtext.sh understand CUDA_VISIBLE_DEVICES #3651
Triggered via pull request
December 16, 2024 21:26
nouiz
Status
Failure
Total duration
17h 10m 13s
Artifacts
84
ci.yaml
on: pull_request
metadata
0s
Matrix: amd64 / test-distribution
Matrix: arm64 / test-distribution
Matrix: amd64 / test-jax / run-unit-test
amd64
/
...
/
launch-slurm-runner
19m 4s
amd64
/
test-nsys-jax-eks
5m 12s
Matrix: amd64 / test-nsys-jax / run-unit-test
Matrix: arm64 / test-jax / run-unit-test
Waiting for pending jobs
arm64
/
test-nsys-jax-eks
0s
arm64
/
...
/
launch-slurm-runner
Matrix: arm64 / test-nsys-jax / run-unit-test
Waiting for pending jobs
Matrix: amd64 / test-te / run-unit-test
Matrix: amd64 / test-upstream-pax / pax-multi-node
Matrix: amd64 / test-upstream-pax / single-process-evaluation
Matrix: amd64 / test-upstream-pax / single-process-multi-device
Matrix: amd64 / test-te-multigpu / te-multi-gpu
Matrix: amd64 / test-upstream-t5x / t5x-multi-gpu
Matrix: amd64 / test-upstream-t5x / t5x-multi-node
Matrix: amd64 / test-gemma / run-unit-test
Matrix: amd64 / test-levanter / run-unit-test
Matrix: amd64 / test-maxtext / maxtext-multinode
Matrix: amd64 / test-maxtext / single-process-multi-device
Matrix: amd64 / test-triton / run-unit-test
Matrix: amd64 / test-nsys-jax-archive
Matrix: arm64 / test-te / run-unit-test
Waiting for pending jobs
Matrix: arm64 / test-upstream-pax / pax-multi-node
Waiting for pending jobs
Matrix: arm64 / test-upstream-pax / single-process-evaluation
Waiting for pending jobs
Matrix: arm64 / test-upstream-pax / single-process-multi-device
Waiting for pending jobs
Matrix: arm64 / test-te-multigpu / te-multi-gpu
Waiting for pending jobs
Matrix: arm64 / test-upstream-t5x / t5x-multi-gpu
Waiting for pending jobs
Matrix: arm64 / test-upstream-t5x / t5x-multi-node
Waiting for pending jobs
Matrix: arm64 / test-gemma / run-unit-test
Waiting for pending jobs
Matrix: arm64 / test-levanter / run-unit-test
Waiting for pending jobs
Matrix: arm64 / test-maxtext / maxtext-multinode
Waiting for pending jobs
Matrix: arm64 / test-maxtext / single-process-multi-device
Waiting for pending jobs
Matrix: arm64 / test-triton / run-unit-test
Waiting for pending jobs
Matrix: arm64 / test-nsys-jax-archive
Matrix: amd64 / test-rosetta-pax / rosetta-pax-multi-node-te
Waiting for pending jobs
Matrix: amd64 / test-rosetta-pax / rosetta-pax-multi-node
Waiting for pending jobs
Matrix: amd64 / test-rosetta-pax / rosetta-pax-single-node-dropout-te
Waiting for pending jobs
Matrix: amd64 / test-rosetta-pax / single-process-evaluation-te
Waiting for pending jobs
Matrix: amd64 / test-rosetta-pax / single-process-multi-device-te
Waiting for pending jobs
Matrix: amd64 / test-rosetta-t5x / multi-gpu-multi-node
Matrix: amd64 / test-rosetta-t5x / single-process-multi-device
Matrix: amd64 / test-rosetta-t5x / vit-multi-gpu-multi-node
Matrix: amd64 / test-rosetta-t5x / vit-single-process-multi-device
Matrix: arm64 / test-rosetta-pax / rosetta-pax-multi-node-te
Waiting for pending jobs
Matrix: arm64 / test-rosetta-pax / rosetta-pax-multi-node
Waiting for pending jobs
Matrix: arm64 / test-rosetta-pax / rosetta-pax-single-node-dropout-te
Waiting for pending jobs
Matrix: arm64 / test-rosetta-pax / single-process-evaluation-te
Waiting for pending jobs
Matrix: arm64 / test-rosetta-pax / single-process-multi-device-te
Waiting for pending jobs
Matrix: arm64 / test-rosetta-t5x / multi-gpu-multi-node
Waiting for pending jobs
Matrix: arm64 / test-rosetta-t5x / single-process-multi-device
Waiting for pending jobs
Matrix: arm64 / test-rosetta-t5x / vit-multi-gpu-multi-node
Waiting for pending jobs
Matrix: arm64 / test-rosetta-t5x / vit-single-process-multi-device
Waiting for pending jobs
Matrix: publish-containers
finalize
/
publish-badge
2s
Annotations
26 errors and 1 warning
|
arm64 / build-rosetta-pax / build-rosetta
buildx failed with: ERROR: failed to solve: process "/bin/sh -c <<\"EOF\" bash -exu\nMANIFEST_DIR=$(dirname ${MANIFEST_FILE})\nif [[ \"${UPDATE_PATCHES}\" != \"true\" && \"${UPDATE_PATCHES}\" != \"false\" ]]; then\n echo \"UPDATE_PATCHES can only be true or false\"\n exit 1\nfi\nif [[ \"${UPDATE_PATCHES}\" == \"true\" ]]; then\n cp -r /mnt/jax-toolbox/.github/container/patches ${MANIFEST_DIR}/\n cp /mnt/jax-toolbox/.github/container/manifest.yaml ${MANIFEST_DIR}/manifest.yaml\n cp /mnt/jax-toolbox/.github/container/create-distribution.sh ${MANIFEST_DIR}/create-distribution.sh\n # TODO: remove\n cp /mnt/jax-toolbox/.github/container/pip-finalize.sh /usr/local/bin/\nfi\ncp -r /mnt/jax-toolbox/rosetta /opt/rosetta\n\nif [[ -n \"${UPDATED_TE_REF}\" ]]; then\n TE_INSTALL_DIR=/opt/transformer-engine\n yq e \".transformer-engine.latest_verified_commit = \\\"${UPDATED_TE_REF}\\\"\" -i $MANIFEST_FILE\n # Install from source instead of pre-built wheel\n sed -i -E 's@( file:///opt/transformer-engine)/dist/[^ ]*@\\1@' /opt/pip-tools.d/requirements-te.in\n git -C $TE_INSTALL_DIR fetch -a\n if [[ \"${UPDATED_TE_REF}\" =~ ^pull/ ]]; then\n PR_ID=$(cut -d/ -f2 <<<\"${UPDATED_TE_REF}\")\n git -C $TE_INSTALL_DIR fetch origin ${UPDATED_TE_REF}:PR-${PR_ID}\n git -C $TE_INSTALL_DIR checkout PR-${PR_ID}\n else\n git -C $TE_INSTALL_DIR checkout ${UPDATED_TE_REF}\n fi\nfi\n\n# Setting the username/email is required to author commits from patches\ngit config --global user.email \"${GIT_USER_EMAIL}\"\ngit config --global user.name \"${GIT_USER_NAME}\"\n\nbash ${MANIFEST_DIR}/create-distribution.sh \\\n --manifest ${MANIFEST_FILE} \\\n --package paxml\nbash ${MANIFEST_DIR}/create-distribution.sh \\\n --manifest ${MANIFEST_FILE} \\\n --package praxis\n# Remove .gitconfig to avoid end-user authoring commits as the \"build user\"\nrm -f ~/.gitconfig\nEOF" did not complete successfully: exit code: 1
|
|
amd64 / build-rosetta-pax / build-rosetta
buildx failed with: ERROR: failed to solve: process "/bin/sh -c <<\"EOF\" bash -exu\nMANIFEST_DIR=$(dirname ${MANIFEST_FILE})\nif [[ \"${UPDATE_PATCHES}\" != \"true\" && \"${UPDATE_PATCHES}\" != \"false\" ]]; then\n echo \"UPDATE_PATCHES can only be true or false\"\n exit 1\nfi\nif [[ \"${UPDATE_PATCHES}\" == \"true\" ]]; then\n cp -r /mnt/jax-toolbox/.github/container/patches ${MANIFEST_DIR}/\n cp /mnt/jax-toolbox/.github/container/manifest.yaml ${MANIFEST_DIR}/manifest.yaml\n cp /mnt/jax-toolbox/.github/container/create-distribution.sh ${MANIFEST_DIR}/create-distribution.sh\n # TODO: remove\n cp /mnt/jax-toolbox/.github/container/pip-finalize.sh /usr/local/bin/\nfi\ncp -r /mnt/jax-toolbox/rosetta /opt/rosetta\n\nif [[ -n \"${UPDATED_TE_REF}\" ]]; then\n TE_INSTALL_DIR=/opt/transformer-engine\n yq e \".transformer-engine.latest_verified_commit = \\\"${UPDATED_TE_REF}\\\"\" -i $MANIFEST_FILE\n # Install from source instead of pre-built wheel\n sed -i -E 's@( file:///opt/transformer-engine)/dist/[^ ]*@\\1@' /opt/pip-tools.d/requirements-te.in\n git -C $TE_INSTALL_DIR fetch -a\n if [[ \"${UPDATED_TE_REF}\" =~ ^pull/ ]]; then\n PR_ID=$(cut -d/ -f2 <<<\"${UPDATED_TE_REF}\")\n git -C $TE_INSTALL_DIR fetch origin ${UPDATED_TE_REF}:PR-${PR_ID}\n git -C $TE_INSTALL_DIR checkout PR-${PR_ID}\n else\n git -C $TE_INSTALL_DIR checkout ${UPDATED_TE_REF}\n fi\nfi\n\n# Setting the username/email is required to author commits from patches\ngit config --global user.email \"${GIT_USER_EMAIL}\"\ngit config --global user.name \"${GIT_USER_NAME}\"\n\nbash ${MANIFEST_DIR}/create-distribution.sh \\\n --manifest ${MANIFEST_FILE} \\\n --package paxml\nbash ${MANIFEST_DIR}/create-distribution.sh \\\n --manifest ${MANIFEST_FILE} \\\n --package praxis\n# Remove .gitconfig to avoid end-user authoring commits as the \"build user\"\nrm -f ~/.gitconfig\nEOF" did not complete successfully: exit code: 1
|
|
amd64 / test-triton / triton-A100-unit-test
Process completed with exit code 1.
|
|
amd64 / test-triton / triton-V100-unit-test
Process completed with exit code 1.
|
|
amd64 / test-upstream-t5x / t5x-multi-node (8G2N, 8, 2)
The job running on runner jumpbox-vc69x-c92lr has exceeded the maximum execution time of 360 minutes.
|
|
amd64 / test-upstream-t5x / t5x-multi-node (8G2N, 8, 2)
The operation was canceled.
|
|
amd64 / test-rosetta-t5x / multi-gpu-multi-node (2N8G-te-1, 8, 2, --gin.train/utils.DatasetConfig.pack=False --gin.train_eva...
The job running on runner jumpbox-vc69x-5ppqf has exceeded the maximum execution time of 360 minutes.
|
|
|
|
amd64 / test-rosetta-t5x / vit-multi-gpu-multi-node (1, 2)
The job running on runner jumpbox-vc69x-c2cs4 has exceeded the maximum execution time of 360 minutes.
|
|
amd64 / test-rosetta-t5x / vit-multi-gpu-multi-node (1, 2)
The operation was canceled.
|
|
amd64 / test-maxtext / maxtext-multinode (1, 4, 2, 2)
The job running on runner jumpbox-vc69x-c4lxg has exceeded the maximum execution time of 360 minutes.
|
|
amd64 / test-maxtext / maxtext-multinode (1, 4, 2, 2)
The operation was canceled.
|
|
amd64 / test-maxtext / test-maxtext-outcome
Process completed with exit code 1.
|
|
amd64 / test-upstream-pax / pax-multi-node (16DP1FSDP1TP1PP, 1, 16, 1, 1, 4)
The job running on runner jumpbox-vc69x-x4lmn has exceeded the maximum execution time of 360 minutes.
|
|
amd64 / test-upstream-pax / pax-multi-node (16DP1FSDP1TP1PP, 1, 16, 1, 1, 4)
The operation was canceled.
|
|
amd64 / test-upstream-t5x / t5x-multi-node (8G2N_fmha, 8, 2, --enable-fmha 1)
The job running on runner jumpbox-vc69x-ww77v has exceeded the maximum execution time of 360 minutes.
|
|
amd64 / test-upstream-t5x / t5x-multi-node (8G2N_fmha, 8, 2, --enable-fmha 1)
The operation was canceled.
|
|
amd64 / test-upstream-t5x / test-upstream-t5x-outcome
Process completed with exit code 1.
|
|
amd64 / test-rosetta-t5x / multi-gpu-multi-node (2N2G_te-0, 2, 2, --enable-te 0)
The job running on runner jumpbox-vc69x-fvplx has exceeded the maximum execution time of 360 minutes.
|
|
amd64 / test-rosetta-t5x / multi-gpu-multi-node (2N2G_te-0, 2, 2, --enable-te 0)
The operation was canceled.
|
|
amd64 / test-rosetta-t5x / vit-multi-gpu-multi-node (8, 2)
The job running on runner jumpbox-vc69x-2tsbp has exceeded the maximum execution time of 360 minutes.
|
|
amd64 / test-rosetta-t5x / vit-multi-gpu-multi-node (8, 2)
The operation was canceled.
|
|
amd64 / test-rosetta-t5x / test-t5x-rosetta-outcome
Process completed with exit code 1.
|
|
amd64 / test-upstream-pax / pax-multi-node (2DP1FSDP2TP4PP, 4, 2, 1, 2, 4)
The job running on runner jumpbox-vc69x-sd4xt has exceeded the maximum execution time of 360 minutes.
|
|
amd64 / test-upstream-pax / pax-multi-node (2DP1FSDP2TP4PP, 4, 2, 1, 2, 4)
The operation was canceled.
|
|
amd64 / test-upstream-pax / test-upstream-pax-outcome
Process completed with exit code 1.
|
|
amd64 / test-te-multigpu / sitrep
ubuntu-latest pipelines will use ubuntu-24.04 soon. For more details, see https://github.com/actions/runner-images/issues/10636
|
Artifacts
Produced during runtime
| Name | Size | |
|---|---|---|
|
artifact-base-build-amd64
|
567 Bytes |
|
|
artifact-base-build-arm64
|
565 Bytes |
|
|
artifact-equinox-build-amd64
|
569 Bytes |
|
|
artifact-equinox-build-arm64
|
569 Bytes |
|
|
artifact-final-report
|
4.48 KB |
|
|
artifact-gemma-build-amd64
|
559 Bytes |
|
|
artifact-jax-build-amd64
|
554 Bytes |
|
|
artifact-jax-build-arm64
|
554 Bytes |
|
|
artifact-levanter-build-amd64
|
572 Bytes |
|
|
artifact-levanter-build-arm64
|
571 Bytes |
|
|
artifact-maxtext-build-amd64
|
566 Bytes |
|
|
artifact-maxtext-build-arm64
|
569 Bytes |
|
|
artifact-maxtext-test
|
654 Bytes |
|
|
artifact-pax-build-amd64
|
570 Bytes |
|
|
artifact-pax-build-arm64
|
567 Bytes |
|
|
artifact-rosetta-build-pax-amd64
|
489 Bytes |
|
|
artifact-rosetta-build-pax-arm64
|
489 Bytes |
|
|
artifact-rosetta-build-t5x-amd64
|
585 Bytes |
|
|
artifact-rosetta-build-t5x-arm64
|
584 Bytes |
|
|
artifact-rosetta-t5x-mgmn-test
|
658 Bytes |
|
|
artifact-t5x-build-amd64
|
568 Bytes |
|
|
artifact-t5x-build-arm64
|
568 Bytes |
|
|
artifact-triton-build-amd64
|
564 Bytes |
|
|
artifact-upstream-pax-mgmn-test
|
687 Bytes |
|
|
artifact-upstream-t5x-mgmn-test
|
634 Bytes |
|
|
artifact-workflow-metadata
|
277 Bytes |
|
|
bumped-manifest
|
81.2 KB |
|
|
final-base
|
254 Bytes |
|
|
final-equinox
|
263 Bytes |
|
|
final-gemma
|
257 Bytes |
|
|
final-jax
|
251 Bytes |
|
|
final-levanter
|
266 Bytes |
|
|
final-maxtext
|
263 Bytes |
|
|
final-t5x
|
251 Bytes |
|
|
final-triton
|
260 Bytes |
|
|
final-upstream-pax
|
277 Bytes |
|
|
final-upstream-t5x
|
276 Bytes |
|
|
gemma-unit-test-A100
|
5.1 KB |
|
|
gemma-unit-test-V100
|
3.83 KB |
|
|
jax-unit-test-A100
|
19.9 KB |
|
|
jax-unit-test-V100
|
26.8 KB |
|
|
levanter-unit-test-A100
|
7.64 KB |
|
|
levanter-unit-test-V100
|
9.57 KB |
|
|
mealkit-equinox
|
272 Bytes |
|
|
mealkit-gemma
|
266 Bytes |
|
|
mealkit-jax
|
261 Bytes |
|
|
mealkit-levanter
|
275 Bytes |
|
|
mealkit-maxtext
|
272 Bytes |
|
|
mealkit-t5x
|
261 Bytes |
|
|
mealkit-triton
|
269 Bytes |
|
|
mealkit-upstream-pax
|
285 Bytes |
|
|
mealkit-upstream-t5x
|
286 Bytes |
|
|
nsys-jax-unit-test-A100
|
29.1 MB |
|
|
nsys-jax-unit-test-V100
|
15.8 MB |
|
|
rosetta-t5x-12361588115-1P1G_te-0
|
35.2 KB |
|
|
rosetta-t5x-12361588115-1P8G_te-1
|
35.3 KB |
|
|
rosetta-t5x-vit-12361588115-VIT1G1N
|
21.6 KB |
|
|
rosetta-t5x-vit-12361588115-VIT1P8G
|
21.6 KB |
|
|
rosetta-t5x-vit-12361588115-VIT8G1N
|
33.1 KB |
|
|
te-2GPU
|
5.64 KB |
|
|
te-4GPU
|
5.69 KB |
|
|
te-8GPU
|
5.71 KB |
|
|
te-multigpu-test
|
529 Bytes |
|
|
te-unit-test-A100
|
773 KB |
|
|
te-unit-test-V100
|
677 KB |
|
|
triton-unit-test-A100
|
3.02 KB |
|
|
triton-unit-test-V100
|
3.07 KB |
|
|
upstream-maxtext-12361588115-1DP1FSDP1TP1PP
|
15.2 KB |
|
|
upstream-maxtext-12361588115-1DP1FSDP8TP1PP
|
21.5 KB |
|
|
upstream-maxtext-12361588115-1DP2FSDP4TP1PP_single_process
|
15.5 KB |
|
|
upstream-maxtext-12361588115-1DP4FSDP2TP1PP
|
21.7 KB |
|
|
upstream-maxtext-12361588115-1DP8FSDP1TP1PP
|
21.5 KB |
|
|
upstream-maxtext-12361588115-2DP2FSDP2TP1PP
|
21.6 KB |
|
|
upstream-pax-12361588115-1DP1FSDP1TP1PP
|
2.66 KB |
|
|
upstream-pax-12361588115-1DP2FSDP4TP1PP_single_process
|
2.75 KB |
|
|
upstream-pax-12361588115-1DP8FSDP1TP1PP
|
5.04 KB |
|
|
upstream-pax-12361588115-2DP1FSDP1TP4PP
|
5.07 KB |
|
|
upstream-pax-12361588115-4DP1FSDP2TP1PP
|
5.12 KB |
|
|
upstream-pax-12361588115-8DP1FSDP1TP1PP
|
5.03 KB |
|
|
upstream-pax-12361588115-8DP1FSDP1TP1PP_eval
|
2.73 KB |
|
|
upstream-pax-12361588115-8DP1FSDP1TP1PP_single_process
|
2.75 KB |
|
|
upstream-pax-12361588115-LLaMA_eval
|
5.36 KB |
|
|
upstream-t5x-12361588115-1P2G_fmha
|
34 KB |
|
|
upstream-t5x-12361588115-1P8G
|
33.6 KB |
|