[ET-VK][CI] Add test-vulkan-genai job for Parakeet on NVIDIA GPU runner#18335
Open
SS-JIA wants to merge 3 commits intogh/SS-JIA/497/basefrom
Open
[ET-VK][CI] Add test-vulkan-genai job for Parakeet on NVIDIA GPU runner#18335SS-JIA wants to merge 3 commits intogh/SS-JIA/497/basefrom
SS-JIA wants to merge 3 commits intogh/SS-JIA/497/basefrom
Conversation
Add a new GitHub CI job that exports and runs the Parakeet TDT model with the Vulkan backend on an NVIDIA GPU runner. The Vulkan export and runner code already exists but had no CI coverage. - Add `--gpu` flag to `setup-vulkan-linux-deps.sh` to skip SwiftShader installation when running on machines with a real GPU driver - Add `vulkan` as a supported device in `export_model_artifact.sh` and `test_model_e2e.sh` - Add `test-vulkan-genai` job to `pull.yml` on `linux.g5.4xlarge.nvidia.gpu` Differential Revision: [D97344728](https://our.internmc.facebook.com/intern/diff/D97344728/) [ghstack-poisoned]
pytorch-bot
bot
added
the
module: vulkan
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/18335
Note: Links to docs will display an error until the docs builds have been completed. ❌ 1 New Failure, 3 Unrelated FailuresAs of commit ef37ef5 with merge base 9076110 ( NEW FAILURE - The following job has failed:
FLAKY - The following jobs failed but were likely due to flakiness present on trunk:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
SS-JIA
pushed a commit
that referenced
this pull request
Mar 19, 2026
Add a new GitHub CI job that exports and runs the Parakeet TDT model with the Vulkan backend on an NVIDIA GPU runner. The Vulkan export and runner code already exists but had no CI coverage. - Add `--gpu` flag to `setup-vulkan-linux-deps.sh` to skip SwiftShader installation when running on machines with a real GPU driver - Add `vulkan` as a supported device in `export_model_artifact.sh` and `test_model_e2e.sh` - Add `test-vulkan-genai` job to `pull.yml` on `linux.g5.4xlarge.nvidia.gpu` Differential Revision: [D97344728](https://our.internmc.facebook.com/intern/diff/D97344728/) ghstack-source-id: 354758467 Pull Request resolved: #18335
meta-cla
bot
added
the
CLA Signed
This PR needs a
|
SS-JIA
temporarily deployed
to
upload-benchmark-results
GitHub Actions
Inactive
…IA GPU runner" Add a new GitHub CI job that exports and runs the Parakeet TDT model with the Vulkan backend on an NVIDIA GPU runner. The Vulkan export and runner code already exists but had no CI coverage. - Add `--gpu` flag to `setup-vulkan-linux-deps.sh` to skip SwiftShader installation when running on machines with a real GPU driver - Add `vulkan` as a supported device in `export_model_artifact.sh` and `test_model_e2e.sh` - Add `test-vulkan-genai` job to `pull.yml` on `linux.g5.4xlarge.nvidia.gpu` Differential Revision: [D97344728](https://our.internmc.facebook.com/intern/diff/D97344728/) cc manuelcandales digantdesai cbilgin [ghstack-poisoned]
SS-JIA
pushed a commit
that referenced
this pull request
Mar 19, 2026
Pull Request resolved: #18335 Add a new GitHub CI job that exports and runs the Parakeet TDT model with the Vulkan backend on an NVIDIA GPU runner. The Vulkan export and runner code already exists but had no CI coverage. - Add `--gpu` flag to `setup-vulkan-linux-deps.sh` to skip SwiftShader installation when running on machines with a real GPU driver - Add `vulkan` as a supported device in `export_model_artifact.sh` and `test_model_e2e.sh` - Add `test-vulkan-genai` job to `pull.yml` on `linux.g5.4xlarge.nvidia.gpu` ghstack-source-id: 354902046 @exported-using-ghexport Differential Revision: [D97344728](https://our.internmc.facebook.com/intern/diff/D97344728/)
SS-JIA
temporarily deployed
to
upload-benchmark-results
GitHub Actions
Inactive
…IA GPU runner" Add a new GitHub CI job that exports and runs the Parakeet TDT model with the Vulkan backend on an NVIDIA GPU runner. The Vulkan export and runner code already exists but had no CI coverage. - Add `--gpu` flag to `setup-vulkan-linux-deps.sh` to skip SwiftShader installation when running on machines with a real GPU driver - Add `vulkan` as a supported device in `export_model_artifact.sh` and `test_model_e2e.sh` - Add `test-vulkan-genai` job to `pull.yml` on `linux.g5.4xlarge.nvidia.gpu` Differential Revision: [D97344728](https://our.internmc.facebook.com/intern/diff/D97344728/) cc manuelcandales digantdesai cbilgin [ghstack-poisoned]
SS-JIA
pushed a commit
that referenced
this pull request
Mar 20, 2026
Pull Request resolved: #18335 Add a new GitHub CI job that exports and runs the Parakeet TDT model with the Vulkan backend on an NVIDIA GPU runner. The Vulkan export and runner code already exists but had no CI coverage. - Add `--gpu` flag to `setup-vulkan-linux-deps.sh` to skip SwiftShader installation when running on machines with a real GPU driver - Add `vulkan` as a supported device in `export_model_artifact.sh` and `test_model_e2e.sh` - Add `test-vulkan-genai` job to `pull.yml` on `linux.g5.4xlarge.nvidia.gpu` - Fix Parakeet CMakeLists.txt to guard `quantized_ops_lib` and `custom_ops` with `if(TARGET ...)` and IMPORTED checks. When the Parakeet runner is built as a standalone CMake project (the second step of `make parakeet-vulkan`), these targets are found via `find_package(executorch)` as imported targets. The existing code called `executorch_target_link_options_shared_lib()` on them unconditionally, which internally calls `target_link_options()` — and CMake does not allow `target_link_options()` on imported targets. Other targets in the same file (e.g. `optimized_native_cpu_ops_lib`, `xnnpack_backend`) already had the correct guards; `quantized_ops_lib` and `custom_ops` were simply missing them. ghstack-source-id: 355090034 @exported-using-ghexport Differential Revision: [D97344728](https://our.internmc.facebook.com/intern/diff/D97344728/)
SS-JIA
temporarily deployed
to
upload-benchmark-results
GitHub Actions
Inactive
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Stack from ghstack (oldest at bottom):
Add a new GitHub CI job that exports and runs the Parakeet TDT model with
the Vulkan backend on an NVIDIA GPU runner. The Vulkan export and runner
code already exists but had no CI coverage.
--gpuflag tosetup-vulkan-linux-deps.shto skip SwiftShaderinstallation when running on machines with a real GPU driver
vulkanas a supported device inexport_model_artifact.shandtest_model_e2e.shtest-vulkan-genaijob topull.ymlonlinux.g5.4xlarge.nvidia.gpuDifferential Revision: D97344728
cc @manuelcandales @digantdesai @cbilgin