| commit | f6564426ce316734a733c337bfca6e56ec2396f6 | [log] [tgz] |
|---|---|---|
| author | Tommy Chiang <ototot@chromium.org> | Tue Apr 08 20:39:55 2025 |
| committer | Chromeos LUCI <chromeos-scoped@luci-project-accounts.iam.gserviceaccount.com> | Fri Apr 25 08:38:05 2025 |
| tree | 5dff470273f9f56aed98fef4c50275fb7cd85446 | |
| parent | 7086dbad48676fc1512261e79bf22a21d1e22597 [diff] |
tflite: mtk_neuron: Add BlockedByPreviousInference test This CL adds a test to ensure that we can send two models at the same time while both models trying to utilize every NPU cores. The expected behavior is that one of the model will need to wait until the other model finish, so the time spent of the later model should be the sum of the first model and itself. This test send two identical models to NPU, and check if the time of the second one is around 2x of the first one. My experiments show that usually it's around 1.89x. BUG=b:374245241 TEST=neuron_delegate_test on navi Change-Id: I123c2e6ca5177f0198e8bafe625f062aba27f4be Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/platform/tflite/+/6437795 Reviewed-by: Shik Chen <shik@chromium.org> Tested-by: Tommy Chiang <ototot@google.com> Commit-Queue: Tommy Chiang <ototot@google.com>
This repository hosts the core ChromeOS TFLite components, enabling on-device machine learning (ODML) workloads accelerated by NPU.
The corresponding ebuild can be found at: tensorflow-9999.ebuild
Patches are stored in the patch/ directory and explicitly listed in WORKSPACE.bazel. A helper script, ./script/patcher.py, is included to facilitate patch management within a TFLite workspace.
The typical workflow:
Eject (Download) TensorFlow Source Code
Download the TensorFlow source code into a local git repository with patches applied as individual commits:
./script/patcher.py eject
This creates a new local git repository at tensorflow/.
Modify the TensorFlow Repository
Make changes to the tensorflow/ repository as needed, following standard git workflows. Optionally, include a PATCH_NAME= tag in commit messages to specify the filename of the corresponding patch.
Seal the Repository
Regenerate the patch files and update the WORKSPACE.bazel file:
./script/patcher.py seal
This updates the patches in the patch/ directory and reflects the changes in WORKSPACE.bazel.
It's preferred to submit changes to upstream TensorFlow first and cherry-pick them as patches. This helps minimize divergence and makes TensorFlow updates easier.