Fix lint #10089
Closed
Fix lint #10089
Conversation
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
I have no idea what this file actually does, but it seems like we are supposed to have this?
…ch#9509) Disable one, fix the other. Testing: built internally
pytorch#9511) I planned to do this everywhere and forgot. Clean it all up, leave a note, enforce the note with visibility. This makes sure everything in buck-land gets ET_USE_THREADPOOL. Test Plan: Profiled run on internal model, no longer seeing parallel_for_no_threadpool
…AME_AS_COMPUTE (pytorch#9613) As the title says, this is mostly a few related find-replaces, plus marking SupportedTensorDtypes::SAME_AS_COMPUTE deprecated.
As the code comment says, these APIs are undergoing development (see e.g. pytorch#9613) and it's pretty inconvenient that they're incidentally committed-to externally. Mark them deprecated so we have the option to drop that commitment in (IIUC) 0.7.
Previous attempt to bump HF transformers version to latest is reverted due to llava model imcompatibility. This PR is to just ensure the CI are able to test `optimum-executorch` with latest version of HF transformers and upcoming `executorch==0.6.0`. Note: This change is purely on CI and only for optimum-executorch, should not affect other models like llava. Co-authored-by: Guang Yang <guangyang@fb.com>
Differential Revision: D69994481 Pull Request resolved: pytorch#8703
### Summary We now have CoreML support out of the box for macOS wheels. Let's test it. ### Test plan CI cc @larryliu0820 @lucylq --------- Co-authored-by: Huy Do <huydhn@gmail.com>
Pull Request resolved: pytorch#9588 TSIA ghstack-source-id: 274222179 @exported-using-ghexport Differential Revision: [D70435293](https://our.internmc.facebook.com/intern/diff/D70435293/)
Pull Request resolved: pytorch#9589 TSIA @pytorchbot label "topic: not user facing" ghstack-source-id: 274222181 @exported-using-ghexport Differential Revision: [D71825480](https://our.internmc.facebook.com/intern/diff/D71825480/)
TSIA @pytorchbot label "topic: not user facing" Differential Revision: [D71825477](https://our.internmc.facebook.com/intern/diff/D71825477/)
TSIA @pytorchbot label "topic: not user facing" Differential Revision: [D71825476](https://our.internmc.facebook.com/intern/diff/D71825476/) [ghstack-poisoned]
## Context
Currently, for the `q_8w_linear` shader, both the texture and the buffer variants use the same global work group and local work group setting.
Specially, the global work group is set to `{out.numel(), 1, 1}` and the local work group is set to `{64, 1, 1}`.
However, I believe this results in a very poor memory re-use for the texture shader. In this configuration:
* Within a work group each invocation will be requesting a different row of A - 64 rows of A requested in total
* All work groups will be requesting the same row of B
* One work group will load 65 unique rows from A and B
Compare this to a local work group size of `{8, 8, 1}`
* Across the work group, 8 rows will be loaded from A and 8 rows will be loaded from B
* One work group will load 16 unique rows total from A and B
Evidently, there is better memory re-use in the latter work group as fewer unique rows are loaded.
## Changes
Modify the `q_8w_linear` shader to use `{8, 8, 1}` local wg if possible. If `M` is small, then instead use `{4, 16, 1}` or `{2, 32, 1}` to reduce the number of inactive invocations.
Differential Revision: [D71706489](https://our.internmc.facebook.com/intern/diff/D71706489/)
[ghstack-poisoned]
…encelength Following previous diff now we can utilize entire kv cache to generate more tokens than max prompt length allowed. Differential Revision: D69073908
This can cuase issues with `disable_global_flags` and internal state of the library, this is something which is set when importing this Differential Revision: [D70402061](https://our.internmc.facebook.com/intern/diff/D70402061/) [ghstack-poisoned]
Differential Revision: D71901449 Pull Request resolved: pytorch#9646
Summary: As title Reviewed By: larryliu0820, kirklandsign Differential Revision: D71577157 Co-authored-by: Digant Desai <digantdesai@meta.com>
Differential Revision: D71901794 Pull Request resolved: pytorch#9668
Differential Revision: D71902542 Pull Request resolved: pytorch#9670
## Context The bencmarks generated by the generated operator benchmarks currently have a high amount of copy overhead: 1. Copy from CPU to staging 2. Copy from staging to GPU Buffer/Image And this is done for both inputs and outputs. Since benchmarks are not correctness tests, copying data in/out is not really necessary especially if the compute shader does not have behaviour dependent on the contents of the input/output tensor. Make it so that by default, the benchmark will only execute the op without adding copy overhead. However, test cases can optionally specify that the copy overhead should be included in the benchmark. Differential Revision: [D71570143](https://our.internmc.facebook.com/intern/diff/D71570143/)
## Context As title; similar to pytorch#9016 since the interface for `ComputePipeline` descriptor was reverted in pytorch#9405. Differential Revision: [D71706868](https://our.internmc.facebook.com/intern/diff/D71706868/) [ghstack-poisoned]
Differential Revision: D71902713 Pull Request resolved: pytorch#9672
…of a shape. Differential Revision: D71903681 Pull Request resolved: pytorch#9673
Differential Revision: D71904351 Pull Request resolved: pytorch#9674
Differential Revision: D71905631 Pull Request resolved: pytorch#9675
Differential Revision: D71905971 Pull Request resolved: pytorch#9676
Differential Revision: D71906972 Pull Request resolved: pytorch#9677
Differential Revision: D71908831 Pull Request resolved: pytorch#9678
Differential Revision: D71909752 Pull Request resolved: pytorch#9679
### Summary Pulling in the non 0.6 changes from: pytorch#10006 pytorch#10016 ### Test plan md
…ch#9355) - Add API to qnn quantizer for setting submodule quant config - Refine QnnQuantizer setting functions --------- Co-authored-by: Chun-I Tsai <chunit@qti.qualcomm.com>
Remove unnecessary line Fix ETDump part
Differential Revision: D72616610 Pull Request resolved: pytorch#9960
Differential Revision: D72440313 Pull Request resolved: pytorch#9894
Differential Revision: D72754398 Pull Request resolved: pytorch#10032
Tests in test_sigmoid_16bit.py and test_sigmoid_32bit.py randomly fails due to a Vela bug regarding handling of the table op sigmoid is converted to. Set all affected tests to flaky until the bug resolved. Co-authored-by: Martin Lindström <Martin.Lindstroem@arm.com>
MobileNetV3 was sporadically failing with the previously set absolute difference threshold. Raise it to prevent flaky test status. Co-authored-by: Martin Lindström <Martin.Lindstroem@arm.com>
Summary: ## Context pytorch#9938 made it so that `linalg_vector_norm` is now decomposed when exporting to Edge. However, this broke some tests in the arm delegate because export passes cannot handle the decomposed operator sequence. To account for this, add `xfail` for the failing tests since `linalg_vector_norm` is not supported in TOSA yet. ## Changes Add `xfail` for `norm` tests in `test_torch_functions.py` Test Plan: ## Test Plan Check CI that failing test is recovered.
Test currently xfails with key error relating to scalar_tensor Signed-off-by: Ryan O'Shea <ryan.oshea3@arm.com>
…dim 1 or 2 (pytorch#10060) For quantized SDPA we want to evaluate performance impact of having seq at dim 1 as well as dim 2. This diff refactors the code to enable this. The same should be done also for float SDPA but left for future. Differential Revision: [D71833060](https://our.internmc.facebook.com/intern/diff/D71833060/)
…torch#10061) Because old name was misnomer ghstack-source-id: 277233486 @exported-using-ghexport Differential Revision: [D71833067](https://our.internmc.facebook.com/intern/diff/D71833067/)
Enable leveraging quantized sdpa op when quantized kv cache is used. Instead of adding yet another arg, at the moment I have chosen to leverage quantize_kv_cache option. Differential Revision: [D71833064](https://our.internmc.facebook.com/intern/diff/D71833064/)
Propagating some changes made to the release/0.6 docs so that future release can get them too
…rch#9266) ### Summary - e2e script for https://github.com/yformer/EfficientSAM - Fastvit breakage fix - Add support for cum_sum - Add bicubic interpolate transform pass - Fix stack op ### Test plan ``` bash python ./examples/qualcomm/oss_scripts/efficientSAM/efficientSAM.py -m ${soc} -b build-android -H ${host_id} -s ${device_id} --oss_repo ${Path_to_oss_repo} --pretrained_weight ${Path_to_pretrained_weight} -d ${Path_to_dataset_dir} ```
… them themselves Differential Revision: D72600295 Pull Request resolved: pytorch#9952
Differential Revision: D72796889 Pull Request resolved: pytorch#10067
…ar_type (pytorch#10076) Following pytorch#9971 - Update get_flatbuffer_scalar_type return type to Result<T> - Iteratively update functions that calling the functions with result type changed: - Check returns, if with an error, pass above the error. - If unable to pass error, update the return type as Result<T> Differential Revision: [D72771753](https://our.internmc.facebook.com/intern/diff/D72771753/)
We have the recipe and .pte file on ExecuTorch-Community on HF. So let's just use that.
…ytorch#10054) Summary: - Support if the rank of input tensor is less than the rank of output tensor. - make_quantizer kwargs alignment. - Remove module.eval() since calling eval() is not supported for exported models. ### Test plan ``` bash python -m backends.qualcomm.tests.test_qnn_delegate TestQNNQuantizedOperator.test_qnn_backend_expand -s ${device_id} -H ${host_id} -m ${soc} -b build-android ```
See error from: pytorch#10063
scripts/build_android_library.sh will no longer build demo app.
…#9989) Add README for Android Make instrumentation test easier for users on their local development.
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/10089
Note: Links to docs will display an error until the docs builds have been completed. This comment was automatically generated by Dr. CI and updates every 15 minutes. |
facebook-github-bot
added
the
CLA Signed
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
Fix lint error from #10054