TFLM compression documentation #2710
Assignees
Comments
ddavis-2015
added a commit
to ddavis-2015/tflite-micro
that referenced
this issue
Oct 2, 2024
Add documentation describing some compression/decompression internals and makefile build procedures. bug=TFLM compression documentation tensorflow#2710
mergify bot
pushed a commit
that referenced
this issue
Oct 3, 2024
@tensorflow/micro Add documentation describing some compression/decompression internals and makefile build procedures. bug=#2710
ddavis-2015
added a commit
to ddavis-2015/tflite-micro
that referenced
this issue
Oct 19, 2024
commit eaee851 Author: ddavis-2015 <[email protected]> Date: Fri Oct 18 17:48:48 2024 -0700 Squashed commit of the following: commit 4894265 Author: ddavis-2015 <[email protected]> Date: Fri Oct 18 17:48:05 2024 -0700 pre-merge empty commit commit a110e41 Author: ddavis-2015 <[email protected]> Date: Fri Oct 18 16:17:13 2024 -0700 fix C++ bitwidth 6 & 7 decompression commit efedcc2 Author: ddavis-2015 <[email protected]> Date: Fri Oct 18 10:18:50 2024 -0700 working decompression unit test commit 81ecf2e Author: ddavis-2015 <[email protected]> Date: Thu Oct 17 18:17:06 2024 -0700 decompression unit test improvements commit b318421 Author: ddavis-2015 <[email protected]> Date: Wed Oct 16 17:34:09 2024 -0700 add decompression unit test commit 9bb2b63 Author: ddavis-2015 <[email protected]> Date: Sun Oct 13 18:34:01 2024 -0700 cleanup commit 77bb05d Author: ddavis-2015 <[email protected]> Date: Sun Oct 13 18:29:33 2024 -0700 align compressed tensor data as per schema commit ad2b1c3 Author: ddavis-2015 <[email protected]> Date: Sat Oct 12 22:35:54 2024 -0700 reduce HIFI5 decompression code size commit 99c6e35 Author: ddavis-2015 <[email protected]> Date: Fri Oct 11 14:02:58 2024 -0700 revert to original Cadence bit width 4 code commit 2388549 Author: ddavis-2015 <[email protected]> Date: Thu Oct 10 17:50:29 2024 -0700 refactor decompression code into reference and platform specific Apply some Xtensa acceleration code changes commit b84853c Author: ddavis-2015 <[email protected]> Date: Tue Oct 8 16:08:55 2024 -0700 testing commit c107f42 Author: Ryan Kuester <[email protected]> Date: Thu Oct 17 14:31:03 2024 -0500 refactor: move misplaced TF_LITE_REMOVE_VIRTUAL_DELETEs to private: Move several TF_LITE_REMOVE_VIRTUAL_DELETE declarations that are wrongly in a public section of their classes. To have the intended effect, as documented in t/l/m/compatibility.h, these must be in a private section. commit 7b3a2bd Author: Ryan Kuester <[email protected]> Date: Thu Oct 17 12:36:46 2024 -0500 build(bazel): always build with TF_LITE_STATIC_MEMORY Add TF_LITE_STATIC_MEMORY to the defines set globally for TFLM builds in Bazel. TFLM always builds with this set in Make, and it appears to have been an oversight that it wasn't set during Bazel builds. Not having it set in Bazel caused some unit tests to pass under Bazel that failed under Make. At the same time, add -fno-exceptions. This flag is also always set in Make builds. Without it, setting TF_LITE_STATIC_MEMORY breaks the build. TF_LITE_STATIC_MEMORY triggers TF_LITE_REMOVE_VIRTUAL_DELETE in t/l/m/compatibility.h, which makes operator delete private in certain classes. When exceptions are enabled, a placement new with those classes is allowed to throw an exception, and operator delete is implicitly called during the unwind. The build breaks because operator delete can't be called if it's private. Disabling exceptions eliminates the unwind code that calls operator delete implicitly, and thus the build succeeds. In any case, -fno-exceptions should have been used in Bazel builds, matching the flags used in Make and the no-exceptions design requirement of the TFLM project. commit 1eb4e0d Author: Ryan Kuester <[email protected]> Date: Thu Oct 17 11:05:45 2024 -0500 feat(python): don't check .sparsity in interpreter Remove the check for sparse tensors in the Python interpreter wrapper. This fixes a broken build when TF_LITE_STATIC_MEMORY is set, which should always be the case in TFLM. TfLiteTensor objects don't have a .sparsity member when TF_LITE_STATIC_MEMORY is set. This prepares for an upcoming commit setting TF_LITE_STATIC_MEMORY during Bazel builds. This hasn't caused build failures in Make builds, which have always set TF_LITE_STATIC_MEMORY, because Make builds don't build the Python interpreter wrapper. commit 7217095 Author: Ryan Kuester <[email protected]> Date: Wed Oct 16 14:03:25 2024 -0500 fix(memory_arena_threshold): with TF_LITE_STATIC_MEMORY Fix the broken build due to redefinition of the threshold when TF_LITE_STATIC_MEMORY is set. Apparently this case isn't triggered in any Bazel test, only in Make. Simplify the threshold specification by only depending on whether compression is enabled and not also on whether TF_LITE_STATIC_MEMORY is in use. commit 8e4e55e Author: Ryan Kuester <[email protected]> Date: Thu Oct 10 12:38:03 2024 -0500 build(bazel): disable codegen when building --//:with_compression The codegen prototype code is not compatible with the changes which implement model compression made to the core TFLM components. For now, disable codegen targets when building with compression enabled. commit 884a234 Author: Ryan Kuester <[email protected]> Date: Tue Oct 15 18:31:01 2024 -0500 build(bazel): compile in compression when --//:with_compression Conditionally compile in support for compressed tensors when the option --//:with_compression is given. commit a1d459b Author: Ryan Kuester <[email protected]> Date: Thu Oct 10 12:28:39 2024 -0500 build(bazel): add --//with_compression build setting Add a --//with_compression user-defined build setting and a corresponding configuration setting. commit 4edc564 Author: Ryan Kuester <[email protected]> Date: Thu Oct 10 12:24:53 2024 -0500 build(bazel): fix compression-related dependencies of micro_allocator commit a52f97f Author: Ryan Kuester <[email protected]> Date: Tue Oct 15 17:28:09 2024 -0500 build(bazel): replace cc_* with tflm_cc_* in remaining TFLM code Replace cc_* targets remaining in TFLM code with tflm_cc_* targets. These are targets which did not formerly use the common copts. Avoid changing imported TFLite code, if for no other reason than to avoid merge conflicts during the automatic sync with upstream TFLite. commit a6368f4 Author: Ryan Kuester <[email protected]> Date: Fri Oct 11 16:08:34 2024 -0500 build(bazel): introduce tflm_cc_* macros, refactoring away micro_copts Remove micro_copts() by replacing every cc_* target that used them with a tflm_cc_* equivalent, and setting those common copts in one place, inside the tflm_cc_* macro. This is the first of several commits introducing tflm_cc_* macros in place of cc_binary, cc_library, and cc_test. Motivated by the upcoming need to support conditional compilation, the objective is to centralize build configuration rather than requiring (and remembering that) each cc_* target in the project add the same common attributes such as compiler options and select()ed #defines. Alternatives such as setting global options on the command line or in .bazelrc, even if simplified with a --config option, fail to preserve flags and hooks for configuration in the case TFLM is used as an external repository by an application project. Nor is it easy in that case for individual targets to override an otherwise global setting. commit 1518422 Author: Ryan Kuester <[email protected]> Date: Thu Oct 10 23:56:49 2024 -0500 chore: remove obsolete ci/temp_patches Remove ci/temp_patches, which was obsoleted in 23f608f once it was no longer used by the sync script. It should have been deleted then. Remove it not only to clean up dead code, but because it contains a reference to `micro_copts`, which is about to be refactored away, and we don't want to leave stray references to it in the tree. commit 18ef080 Author: Ryan Kuester <[email protected]> Date: Tue Oct 8 17:58:12 2024 -0500 refactor: use metadata_saved.h instead of metadata_generated.h Use the generated file metadata_saved.h instead of metadata_generated.h for the reasons explained in t/l/m/compression/BUILD:metadata_saved. Delete metadata_generated.h from the source tree as it is not maintained. commit 5a02e30 Author: Ryan Kuester <[email protected]> Date: Thu Oct 10 13:46:46 2024 -0500 test(memory_arena_threshold): adjust expected value with compression Fix a test failure by setting a different expected value for the persistent buffer allocation when compression is configured in. The allocation was allowed to vary by 3%; however, compression adds ~10%. Set the expected value to the measured value when compression is configured in. commit 01bc582 Author: Ryan Kuester <[email protected]> Date: Thu Oct 10 13:35:10 2024 -0500 test(memory_arena_threshold): don't expect exact allocation values Remove the check for allocation sizes to exactly match expected values. This check immediately followed--and thus rendered pointless---a check that sizes are within a certain percentage, which seems to be the true intent of the test. commit e0aae77 Merge: e328029 e86d97b Author: Ryan Kuester <[email protected]> Date: Wed Oct 16 13:39:56 2024 -0500 Merge branch 'main' into compress-testing commit e328029 Author: Ryan Kuester <[email protected]> Date: Mon Oct 7 12:52:23 2024 -0500 build(bazel): fix dependencies in work-in-progress compression code In the Bazel build, add dependencies needed by the code added to t/l/m:micro_context for decompression. The Bazel build with or without compression was broken without this. commit e86d97b Author: RJ Ascani <[email protected]> Date: Mon Oct 7 10:36:26 2024 -0700 Replace rascani with suleshahid on OWNERS (tensorflow#2715) BUG=none commit b773428 Author: Ryan Kuester <[email protected]> Date: Fri Oct 4 09:59:10 2024 -0500 feat(compression): add work-in-progress compression and viewer tools commit f6bd486 Merge: 487c17a e3f6dc1 Author: Ryan Kuester <[email protected]> Date: Fri Oct 4 09:36:24 2024 -0500 Merge branch 'main' into compress-prerelease commit e3f6dc1 Author: David Davis <[email protected]> Date: Thu Oct 3 10:45:00 2024 -0700 Compression documentation (tensorflow#2711) @tensorflow/micro Add documentation describing some compression/decompression internals and makefile build procedures. bug=tensorflow#2710 commit b3967a9 Author: Ryan Kuester <[email protected]> Date: Wed Oct 2 13:36:01 2024 -0500 style: add .style.yapf to control yapf styling of Python code (tensorflow#2709) Add a .style.yapf file so yapf can be used to style Python code without passing the project's style via command line option. Remove the corresponding patch to pigweed's call to yapf, used by CI, and instead let it too rely on .style.yapf. Remove the developer documentation's instruction to use the command line option. BUG=description commit d249577 Author: Ryan Kuester <[email protected]> Date: Tue Oct 1 16:16:45 2024 -0500 build(codegen): suppress noise in console output (tensorflow#2708) Add a --quiet option to the code_generator binary so that when it's used within the build system, it doesn't print unexpected, distracting noise to the console. Generally, compiler or generator commands don't print output unless there's an error. BUG=description
ddavis-2015
added a commit
to rkuester/tflite-micro
that referenced
this issue
Oct 21, 2024
commit b43c16c Author: ddavis-2015 <[email protected]> Date: Mon Oct 21 05:05:15 2024 -0700 fix code style errors. commit 2d825e3 Author: ddavis-2015 <[email protected]> Date: Mon Oct 21 04:09:39 2024 -0700 fix code style errors. fix BUILD file style errors. commit 459569a Author: ddavis-2015 <[email protected]> Date: Sun Oct 20 19:17:43 2024 -0700 use kernel optimzer level -O3 and -LNO:simd for Xtensa HIFI5 commit 122db20 Author: ddavis-2015 <[email protected]> Date: Sun Oct 20 19:07:19 2024 -0700 add compression build/test to bazel default test script commit d96b614 Author: ddavis-2015 <[email protected]> Date: Sun Oct 20 18:56:31 2024 -0700 fix CI code style errors fix CI BUILD file style errors commit 821dfdf Author: ddavis-2015 <[email protected]> Date: Sun Oct 20 12:19:34 2024 -0700 Cleanup header file usage. Add decompression code to Bazel BUILD files commit 4a02b22 Author: ddavis-2015 <[email protected]> Date: Sat Oct 19 13:31:11 2024 -0700 cleanup commit 3d765e6 Author: ddavis-2015 <[email protected]> Date: Fri Oct 18 18:01:21 2024 -0700 Squashed commit of the following: commit eaee851 Author: ddavis-2015 <[email protected]> Date: Fri Oct 18 17:48:48 2024 -0700 Squashed commit of the following: commit 4894265 Author: ddavis-2015 <[email protected]> Date: Fri Oct 18 17:48:05 2024 -0700 pre-merge empty commit commit a110e41 Author: ddavis-2015 <[email protected]> Date: Fri Oct 18 16:17:13 2024 -0700 fix C++ bitwidth 6 & 7 decompression commit efedcc2 Author: ddavis-2015 <[email protected]> Date: Fri Oct 18 10:18:50 2024 -0700 working decompression unit test commit 81ecf2e Author: ddavis-2015 <[email protected]> Date: Thu Oct 17 18:17:06 2024 -0700 decompression unit test improvements commit b318421 Author: ddavis-2015 <[email protected]> Date: Wed Oct 16 17:34:09 2024 -0700 add decompression unit test commit 9bb2b63 Author: ddavis-2015 <[email protected]> Date: Sun Oct 13 18:34:01 2024 -0700 cleanup commit 77bb05d Author: ddavis-2015 <[email protected]> Date: Sun Oct 13 18:29:33 2024 -0700 align compressed tensor data as per schema commit ad2b1c3 Author: ddavis-2015 <[email protected]> Date: Sat Oct 12 22:35:54 2024 -0700 reduce HIFI5 decompression code size commit 99c6e35 Author: ddavis-2015 <[email protected]> Date: Fri Oct 11 14:02:58 2024 -0700 revert to original Cadence bit width 4 code commit 2388549 Author: ddavis-2015 <[email protected]> Date: Thu Oct 10 17:50:29 2024 -0700 refactor decompression code into reference and platform specific Apply some Xtensa acceleration code changes commit b84853c Author: ddavis-2015 <[email protected]> Date: Tue Oct 8 16:08:55 2024 -0700 testing commit c107f42 Author: Ryan Kuester <[email protected]> Date: Thu Oct 17 14:31:03 2024 -0500 refactor: move misplaced TF_LITE_REMOVE_VIRTUAL_DELETEs to private: Move several TF_LITE_REMOVE_VIRTUAL_DELETE declarations that are wrongly in a public section of their classes. To have the intended effect, as documented in t/l/m/compatibility.h, these must be in a private section. commit 7b3a2bd Author: Ryan Kuester <[email protected]> Date: Thu Oct 17 12:36:46 2024 -0500 build(bazel): always build with TF_LITE_STATIC_MEMORY Add TF_LITE_STATIC_MEMORY to the defines set globally for TFLM builds in Bazel. TFLM always builds with this set in Make, and it appears to have been an oversight that it wasn't set during Bazel builds. Not having it set in Bazel caused some unit tests to pass under Bazel that failed under Make. At the same time, add -fno-exceptions. This flag is also always set in Make builds. Without it, setting TF_LITE_STATIC_MEMORY breaks the build. TF_LITE_STATIC_MEMORY triggers TF_LITE_REMOVE_VIRTUAL_DELETE in t/l/m/compatibility.h, which makes operator delete private in certain classes. When exceptions are enabled, a placement new with those classes is allowed to throw an exception, and operator delete is implicitly called during the unwind. The build breaks because operator delete can't be called if it's private. Disabling exceptions eliminates the unwind code that calls operator delete implicitly, and thus the build succeeds. In any case, -fno-exceptions should have been used in Bazel builds, matching the flags used in Make and the no-exceptions design requirement of the TFLM project. commit 1eb4e0d Author: Ryan Kuester <[email protected]> Date: Thu Oct 17 11:05:45 2024 -0500 feat(python): don't check .sparsity in interpreter Remove the check for sparse tensors in the Python interpreter wrapper. This fixes a broken build when TF_LITE_STATIC_MEMORY is set, which should always be the case in TFLM. TfLiteTensor objects don't have a .sparsity member when TF_LITE_STATIC_MEMORY is set. This prepares for an upcoming commit setting TF_LITE_STATIC_MEMORY during Bazel builds. This hasn't caused build failures in Make builds, which have always set TF_LITE_STATIC_MEMORY, because Make builds don't build the Python interpreter wrapper. commit 7217095 Author: Ryan Kuester <[email protected]> Date: Wed Oct 16 14:03:25 2024 -0500 fix(memory_arena_threshold): with TF_LITE_STATIC_MEMORY Fix the broken build due to redefinition of the threshold when TF_LITE_STATIC_MEMORY is set. Apparently this case isn't triggered in any Bazel test, only in Make. Simplify the threshold specification by only depending on whether compression is enabled and not also on whether TF_LITE_STATIC_MEMORY is in use. commit 8e4e55e Author: Ryan Kuester <[email protected]> Date: Thu Oct 10 12:38:03 2024 -0500 build(bazel): disable codegen when building --//:with_compression The codegen prototype code is not compatible with the changes which implement model compression made to the core TFLM components. For now, disable codegen targets when building with compression enabled. commit 884a234 Author: Ryan Kuester <[email protected]> Date: Tue Oct 15 18:31:01 2024 -0500 build(bazel): compile in compression when --//:with_compression Conditionally compile in support for compressed tensors when the option --//:with_compression is given. commit a1d459b Author: Ryan Kuester <[email protected]> Date: Thu Oct 10 12:28:39 2024 -0500 build(bazel): add --//with_compression build setting Add a --//with_compression user-defined build setting and a corresponding configuration setting. commit 4edc564 Author: Ryan Kuester <[email protected]> Date: Thu Oct 10 12:24:53 2024 -0500 build(bazel): fix compression-related dependencies of micro_allocator commit a52f97f Author: Ryan Kuester <[email protected]> Date: Tue Oct 15 17:28:09 2024 -0500 build(bazel): replace cc_* with tflm_cc_* in remaining TFLM code Replace cc_* targets remaining in TFLM code with tflm_cc_* targets. These are targets which did not formerly use the common copts. Avoid changing imported TFLite code, if for no other reason than to avoid merge conflicts during the automatic sync with upstream TFLite. commit a6368f4 Author: Ryan Kuester <[email protected]> Date: Fri Oct 11 16:08:34 2024 -0500 build(bazel): introduce tflm_cc_* macros, refactoring away micro_copts Remove micro_copts() by replacing every cc_* target that used them with a tflm_cc_* equivalent, and setting those common copts in one place, inside the tflm_cc_* macro. This is the first of several commits introducing tflm_cc_* macros in place of cc_binary, cc_library, and cc_test. Motivated by the upcoming need to support conditional compilation, the objective is to centralize build configuration rather than requiring (and remembering that) each cc_* target in the project add the same common attributes such as compiler options and select()ed #defines. Alternatives such as setting global options on the command line or in .bazelrc, even if simplified with a --config option, fail to preserve flags and hooks for configuration in the case TFLM is used as an external repository by an application project. Nor is it easy in that case for individual targets to override an otherwise global setting. commit 1518422 Author: Ryan Kuester <[email protected]> Date: Thu Oct 10 23:56:49 2024 -0500 chore: remove obsolete ci/temp_patches Remove ci/temp_patches, which was obsoleted in 23f608f once it was no longer used by the sync script. It should have been deleted then. Remove it not only to clean up dead code, but because it contains a reference to `micro_copts`, which is about to be refactored away, and we don't want to leave stray references to it in the tree. commit 18ef080 Author: Ryan Kuester <[email protected]> Date: Tue Oct 8 17:58:12 2024 -0500 refactor: use metadata_saved.h instead of metadata_generated.h Use the generated file metadata_saved.h instead of metadata_generated.h for the reasons explained in t/l/m/compression/BUILD:metadata_saved. Delete metadata_generated.h from the source tree as it is not maintained. commit 5a02e30 Author: Ryan Kuester <[email protected]> Date: Thu Oct 10 13:46:46 2024 -0500 test(memory_arena_threshold): adjust expected value with compression Fix a test failure by setting a different expected value for the persistent buffer allocation when compression is configured in. The allocation was allowed to vary by 3%; however, compression adds ~10%. Set the expected value to the measured value when compression is configured in. commit 01bc582 Author: Ryan Kuester <[email protected]> Date: Thu Oct 10 13:35:10 2024 -0500 test(memory_arena_threshold): don't expect exact allocation values Remove the check for allocation sizes to exactly match expected values. This check immediately followed--and thus rendered pointless---a check that sizes are within a certain percentage, which seems to be the true intent of the test. commit e0aae77 Merge: e328029 e86d97b Author: Ryan Kuester <[email protected]> Date: Wed Oct 16 13:39:56 2024 -0500 Merge branch 'main' into compress-testing commit e328029 Author: Ryan Kuester <[email protected]> Date: Mon Oct 7 12:52:23 2024 -0500 build(bazel): fix dependencies in work-in-progress compression code In the Bazel build, add dependencies needed by the code added to t/l/m:micro_context for decompression. The Bazel build with or without compression was broken without this. commit e86d97b Author: RJ Ascani <[email protected]> Date: Mon Oct 7 10:36:26 2024 -0700 Replace rascani with suleshahid on OWNERS (tensorflow#2715) BUG=none commit b773428 Author: Ryan Kuester <[email protected]> Date: Fri Oct 4 09:59:10 2024 -0500 feat(compression): add work-in-progress compression and viewer tools commit f6bd486 Merge: 487c17a e3f6dc1 Author: Ryan Kuester <[email protected]> Date: Fri Oct 4 09:36:24 2024 -0500 Merge branch 'main' into compress-prerelease commit e3f6dc1 Author: David Davis <[email protected]> Date: Thu Oct 3 10:45:00 2024 -0700 Compression documentation (tensorflow#2711) @tensorflow/micro Add documentation describing some compression/decompression internals and makefile build procedures. bug=tensorflow#2710 commit b3967a9 Author: Ryan Kuester <[email protected]> Date: Wed Oct 2 13:36:01 2024 -0500 style: add .style.yapf to control yapf styling of Python code (tensorflow#2709) Add a .style.yapf file so yapf can be used to style Python code without passing the project's style via command line option. Remove the corresponding patch to pigweed's call to yapf, used by CI, and instead let it too rely on .style.yapf. Remove the developer documentation's instruction to use the command line option. BUG=description commit d249577 Author: Ryan Kuester <[email protected]> Date: Tue Oct 1 16:16:45 2024 -0500 build(codegen): suppress noise in console output (tensorflow#2708) Add a --quiet option to the code_generator binary so that when it's used within the build system, it doesn't print unexpected, distracting noise to the console. Generally, compiler or generator commands don't print output unless there's an error. BUG=description commit 4894265 Author: ddavis-2015 <[email protected]> Date: Fri Oct 18 17:48:05 2024 -0700 pre-merge empty commit commit a110e41 Author: ddavis-2015 <[email protected]> Date: Fri Oct 18 16:17:13 2024 -0700 fix C++ bitwidth 6 & 7 decompression commit efedcc2 Author: ddavis-2015 <[email protected]> Date: Fri Oct 18 10:18:50 2024 -0700 working decompression unit test commit 81ecf2e Author: ddavis-2015 <[email protected]> Date: Thu Oct 17 18:17:06 2024 -0700 decompression unit test improvements commit b318421 Author: ddavis-2015 <[email protected]> Date: Wed Oct 16 17:34:09 2024 -0700 add decompression unit test commit 9bb2b63 Author: ddavis-2015 <[email protected]> Date: Sun Oct 13 18:34:01 2024 -0700 cleanup commit 77bb05d Author: ddavis-2015 <[email protected]> Date: Sun Oct 13 18:29:33 2024 -0700 align compressed tensor data as per schema commit ad2b1c3 Author: ddavis-2015 <[email protected]> Date: Sat Oct 12 22:35:54 2024 -0700 reduce HIFI5 decompression code size commit 99c6e35 Author: ddavis-2015 <[email protected]> Date: Fri Oct 11 14:02:58 2024 -0700 revert to original Cadence bit width 4 code commit 2388549 Author: ddavis-2015 <[email protected]> Date: Thu Oct 10 17:50:29 2024 -0700 refactor decompression code into reference and platform specific Apply some Xtensa acceleration code changes commit b84853c Author: ddavis-2015 <[email protected]> Date: Tue Oct 8 16:08:55 2024 -0700 testing
ddavis-2015
added a commit
to rkuester/tflite-micro
that referenced
this issue
Oct 23, 2024
commit df29a4c Author: ddavis-2015 <[email protected]> Date: Wed Oct 23 03:47:27 2024 -0700 header file cleanup. commit 7dc34a9 Author: ddavis-2015 <[email protected]> Date: Mon Oct 21 17:16:24 2024 -0700 update to latest Cadence decompression code. commit b43c16c Author: ddavis-2015 <[email protected]> Date: Mon Oct 21 05:05:15 2024 -0700 fix code style errors. commit 2d825e3 Author: ddavis-2015 <[email protected]> Date: Mon Oct 21 04:09:39 2024 -0700 fix code style errors. fix BUILD file style errors. commit 459569a Author: ddavis-2015 <[email protected]> Date: Sun Oct 20 19:17:43 2024 -0700 use kernel optimzer level -O3 and -LNO:simd for Xtensa HIFI5 commit 122db20 Author: ddavis-2015 <[email protected]> Date: Sun Oct 20 19:07:19 2024 -0700 add compression build/test to bazel default test script commit d96b614 Author: ddavis-2015 <[email protected]> Date: Sun Oct 20 18:56:31 2024 -0700 fix CI code style errors fix CI BUILD file style errors commit 821dfdf Author: ddavis-2015 <[email protected]> Date: Sun Oct 20 12:19:34 2024 -0700 Cleanup header file usage. Add decompression code to Bazel BUILD files commit 4a02b22 Author: ddavis-2015 <[email protected]> Date: Sat Oct 19 13:31:11 2024 -0700 cleanup commit 3d765e6 Author: ddavis-2015 <[email protected]> Date: Fri Oct 18 18:01:21 2024 -0700 Squashed commit of the following: commit eaee851 Author: ddavis-2015 <[email protected]> Date: Fri Oct 18 17:48:48 2024 -0700 Squashed commit of the following: commit 4894265 Author: ddavis-2015 <[email protected]> Date: Fri Oct 18 17:48:05 2024 -0700 pre-merge empty commit commit a110e41 Author: ddavis-2015 <[email protected]> Date: Fri Oct 18 16:17:13 2024 -0700 fix C++ bitwidth 6 & 7 decompression commit efedcc2 Author: ddavis-2015 <[email protected]> Date: Fri Oct 18 10:18:50 2024 -0700 working decompression unit test commit 81ecf2e Author: ddavis-2015 <[email protected]> Date: Thu Oct 17 18:17:06 2024 -0700 decompression unit test improvements commit b318421 Author: ddavis-2015 <[email protected]> Date: Wed Oct 16 17:34:09 2024 -0700 add decompression unit test commit 9bb2b63 Author: ddavis-2015 <[email protected]> Date: Sun Oct 13 18:34:01 2024 -0700 cleanup commit 77bb05d Author: ddavis-2015 <[email protected]> Date: Sun Oct 13 18:29:33 2024 -0700 align compressed tensor data as per schema commit ad2b1c3 Author: ddavis-2015 <[email protected]> Date: Sat Oct 12 22:35:54 2024 -0700 reduce HIFI5 decompression code size commit 99c6e35 Author: ddavis-2015 <[email protected]> Date: Fri Oct 11 14:02:58 2024 -0700 revert to original Cadence bit width 4 code commit 2388549 Author: ddavis-2015 <[email protected]> Date: Thu Oct 10 17:50:29 2024 -0700 refactor decompression code into reference and platform specific Apply some Xtensa acceleration code changes commit b84853c Author: ddavis-2015 <[email protected]> Date: Tue Oct 8 16:08:55 2024 -0700 testing commit c107f42 Author: Ryan Kuester <[email protected]> Date: Thu Oct 17 14:31:03 2024 -0500 refactor: move misplaced TF_LITE_REMOVE_VIRTUAL_DELETEs to private: Move several TF_LITE_REMOVE_VIRTUAL_DELETE declarations that are wrongly in a public section of their classes. To have the intended effect, as documented in t/l/m/compatibility.h, these must be in a private section. commit 7b3a2bd Author: Ryan Kuester <[email protected]> Date: Thu Oct 17 12:36:46 2024 -0500 build(bazel): always build with TF_LITE_STATIC_MEMORY Add TF_LITE_STATIC_MEMORY to the defines set globally for TFLM builds in Bazel. TFLM always builds with this set in Make, and it appears to have been an oversight that it wasn't set during Bazel builds. Not having it set in Bazel caused some unit tests to pass under Bazel that failed under Make. At the same time, add -fno-exceptions. This flag is also always set in Make builds. Without it, setting TF_LITE_STATIC_MEMORY breaks the build. TF_LITE_STATIC_MEMORY triggers TF_LITE_REMOVE_VIRTUAL_DELETE in t/l/m/compatibility.h, which makes operator delete private in certain classes. When exceptions are enabled, a placement new with those classes is allowed to throw an exception, and operator delete is implicitly called during the unwind. The build breaks because operator delete can't be called if it's private. Disabling exceptions eliminates the unwind code that calls operator delete implicitly, and thus the build succeeds. In any case, -fno-exceptions should have been used in Bazel builds, matching the flags used in Make and the no-exceptions design requirement of the TFLM project. commit 1eb4e0d Author: Ryan Kuester <[email protected]> Date: Thu Oct 17 11:05:45 2024 -0500 feat(python): don't check .sparsity in interpreter Remove the check for sparse tensors in the Python interpreter wrapper. This fixes a broken build when TF_LITE_STATIC_MEMORY is set, which should always be the case in TFLM. TfLiteTensor objects don't have a .sparsity member when TF_LITE_STATIC_MEMORY is set. This prepares for an upcoming commit setting TF_LITE_STATIC_MEMORY during Bazel builds. This hasn't caused build failures in Make builds, which have always set TF_LITE_STATIC_MEMORY, because Make builds don't build the Python interpreter wrapper. commit 7217095 Author: Ryan Kuester <[email protected]> Date: Wed Oct 16 14:03:25 2024 -0500 fix(memory_arena_threshold): with TF_LITE_STATIC_MEMORY Fix the broken build due to redefinition of the threshold when TF_LITE_STATIC_MEMORY is set. Apparently this case isn't triggered in any Bazel test, only in Make. Simplify the threshold specification by only depending on whether compression is enabled and not also on whether TF_LITE_STATIC_MEMORY is in use. commit 8e4e55e Author: Ryan Kuester <[email protected]> Date: Thu Oct 10 12:38:03 2024 -0500 build(bazel): disable codegen when building --//:with_compression The codegen prototype code is not compatible with the changes which implement model compression made to the core TFLM components. For now, disable codegen targets when building with compression enabled. commit 884a234 Author: Ryan Kuester <[email protected]> Date: Tue Oct 15 18:31:01 2024 -0500 build(bazel): compile in compression when --//:with_compression Conditionally compile in support for compressed tensors when the option --//:with_compression is given. commit a1d459b Author: Ryan Kuester <[email protected]> Date: Thu Oct 10 12:28:39 2024 -0500 build(bazel): add --//with_compression build setting Add a --//with_compression user-defined build setting and a corresponding configuration setting. commit 4edc564 Author: Ryan Kuester <[email protected]> Date: Thu Oct 10 12:24:53 2024 -0500 build(bazel): fix compression-related dependencies of micro_allocator commit a52f97f Author: Ryan Kuester <[email protected]> Date: Tue Oct 15 17:28:09 2024 -0500 build(bazel): replace cc_* with tflm_cc_* in remaining TFLM code Replace cc_* targets remaining in TFLM code with tflm_cc_* targets. These are targets which did not formerly use the common copts. Avoid changing imported TFLite code, if for no other reason than to avoid merge conflicts during the automatic sync with upstream TFLite. commit a6368f4 Author: Ryan Kuester <[email protected]> Date: Fri Oct 11 16:08:34 2024 -0500 build(bazel): introduce tflm_cc_* macros, refactoring away micro_copts Remove micro_copts() by replacing every cc_* target that used them with a tflm_cc_* equivalent, and setting those common copts in one place, inside the tflm_cc_* macro. This is the first of several commits introducing tflm_cc_* macros in place of cc_binary, cc_library, and cc_test. Motivated by the upcoming need to support conditional compilation, the objective is to centralize build configuration rather than requiring (and remembering that) each cc_* target in the project add the same common attributes such as compiler options and select()ed #defines. Alternatives such as setting global options on the command line or in .bazelrc, even if simplified with a --config option, fail to preserve flags and hooks for configuration in the case TFLM is used as an external repository by an application project. Nor is it easy in that case for individual targets to override an otherwise global setting. commit 1518422 Author: Ryan Kuester <[email protected]> Date: Thu Oct 10 23:56:49 2024 -0500 chore: remove obsolete ci/temp_patches Remove ci/temp_patches, which was obsoleted in 23f608f once it was no longer used by the sync script. It should have been deleted then. Remove it not only to clean up dead code, but because it contains a reference to `micro_copts`, which is about to be refactored away, and we don't want to leave stray references to it in the tree. commit 18ef080 Author: Ryan Kuester <[email protected]> Date: Tue Oct 8 17:58:12 2024 -0500 refactor: use metadata_saved.h instead of metadata_generated.h Use the generated file metadata_saved.h instead of metadata_generated.h for the reasons explained in t/l/m/compression/BUILD:metadata_saved. Delete metadata_generated.h from the source tree as it is not maintained. commit 5a02e30 Author: Ryan Kuester <[email protected]> Date: Thu Oct 10 13:46:46 2024 -0500 test(memory_arena_threshold): adjust expected value with compression Fix a test failure by setting a different expected value for the persistent buffer allocation when compression is configured in. The allocation was allowed to vary by 3%; however, compression adds ~10%. Set the expected value to the measured value when compression is configured in. commit 01bc582 Author: Ryan Kuester <[email protected]> Date: Thu Oct 10 13:35:10 2024 -0500 test(memory_arena_threshold): don't expect exact allocation values Remove the check for allocation sizes to exactly match expected values. This check immediately followed--and thus rendered pointless---a check that sizes are within a certain percentage, which seems to be the true intent of the test. commit e0aae77 Merge: e328029 e86d97b Author: Ryan Kuester <[email protected]> Date: Wed Oct 16 13:39:56 2024 -0500 Merge branch 'main' into compress-testing commit e328029 Author: Ryan Kuester <[email protected]> Date: Mon Oct 7 12:52:23 2024 -0500 build(bazel): fix dependencies in work-in-progress compression code In the Bazel build, add dependencies needed by the code added to t/l/m:micro_context for decompression. The Bazel build with or without compression was broken without this. commit e86d97b Author: RJ Ascani <[email protected]> Date: Mon Oct 7 10:36:26 2024 -0700 Replace rascani with suleshahid on OWNERS (tensorflow#2715) BUG=none commit b773428 Author: Ryan Kuester <[email protected]> Date: Fri Oct 4 09:59:10 2024 -0500 feat(compression): add work-in-progress compression and viewer tools commit f6bd486 Merge: 487c17a e3f6dc1 Author: Ryan Kuester <[email protected]> Date: Fri Oct 4 09:36:24 2024 -0500 Merge branch 'main' into compress-prerelease commit e3f6dc1 Author: David Davis <[email protected]> Date: Thu Oct 3 10:45:00 2024 -0700 Compression documentation (tensorflow#2711) @tensorflow/micro Add documentation describing some compression/decompression internals and makefile build procedures. bug=tensorflow#2710 commit b3967a9 Author: Ryan Kuester <[email protected]> Date: Wed Oct 2 13:36:01 2024 -0500 style: add .style.yapf to control yapf styling of Python code (tensorflow#2709) Add a .style.yapf file so yapf can be used to style Python code without passing the project's style via command line option. Remove the corresponding patch to pigweed's call to yapf, used by CI, and instead let it too rely on .style.yapf. Remove the developer documentation's instruction to use the command line option. BUG=description commit d249577 Author: Ryan Kuester <[email protected]> Date: Tue Oct 1 16:16:45 2024 -0500 build(codegen): suppress noise in console output (tensorflow#2708) Add a --quiet option to the code_generator binary so that when it's used within the build system, it doesn't print unexpected, distracting noise to the console. Generally, compiler or generator commands don't print output unless there's an error. BUG=description commit 4894265 Author: ddavis-2015 <[email protected]> Date: Fri Oct 18 17:48:05 2024 -0700 pre-merge empty commit commit a110e41 Author: ddavis-2015 <[email protected]> Date: Fri Oct 18 16:17:13 2024 -0700 fix C++ bitwidth 6 & 7 decompression commit efedcc2 Author: ddavis-2015 <[email protected]> Date: Fri Oct 18 10:18:50 2024 -0700 working decompression unit test commit 81ecf2e Author: ddavis-2015 <[email protected]> Date: Thu Oct 17 18:17:06 2024 -0700 decompression unit test improvements commit b318421 Author: ddavis-2015 <[email protected]> Date: Wed Oct 16 17:34:09 2024 -0700 add decompression unit test commit 9bb2b63 Author: ddavis-2015 <[email protected]> Date: Sun Oct 13 18:34:01 2024 -0700 cleanup commit 77bb05d Author: ddavis-2015 <[email protected]> Date: Sun Oct 13 18:29:33 2024 -0700 align compressed tensor data as per schema commit ad2b1c3 Author: ddavis-2015 <[email protected]> Date: Sat Oct 12 22:35:54 2024 -0700 reduce HIFI5 decompression code size commit 99c6e35 Author: ddavis-2015 <[email protected]> Date: Fri Oct 11 14:02:58 2024 -0700 revert to original Cadence bit width 4 code commit 2388549 Author: ddavis-2015 <[email protected]> Date: Thu Oct 10 17:50:29 2024 -0700 refactor decompression code into reference and platform specific Apply some Xtensa acceleration code changes commit b84853c Author: ddavis-2015 <[email protected]> Date: Tue Oct 8 16:08:55 2024 -0700 testing
|
"This issue is being marked as stale due to inactivity. Remove label or comment to prevent closure in 5 days." |
ddavis-2015
added a commit
to rkuester/tflite-micro
that referenced
this issue
Nov 1, 2024
commit 4dca8e7 Author: ddavis-2015 <[email protected]> Date: Thu Oct 31 02:10:17 2024 -0700 keep generic benchmark application binary size stable regardless of whether compression is enabled commit ac4a0a4 Author: ddavis-2015 <[email protected]> Date: Thu Oct 31 01:49:40 2024 -0700 cleanup commit fc4b473 Author: ddavis-2015 <[email protected]> Date: Wed Oct 30 04:19:07 2024 -0700 additional operator relocation check. commit 8c53ee3 Author: ddavis-2015 <[email protected]> Date: Tue Oct 29 17:17:11 2024 -0700 first cut at op relocation script. expand model_facade. fix greedy memory planner debug output. commit df29a4c Author: ddavis-2015 <[email protected]> Date: Wed Oct 23 03:47:27 2024 -0700 header file cleanup. commit 7dc34a9 Author: ddavis-2015 <[email protected]> Date: Mon Oct 21 17:16:24 2024 -0700 update to latest Cadence decompression code. commit b43c16c Author: ddavis-2015 <[email protected]> Date: Mon Oct 21 05:05:15 2024 -0700 fix code style errors. commit 2d825e3 Author: ddavis-2015 <[email protected]> Date: Mon Oct 21 04:09:39 2024 -0700 fix code style errors. fix BUILD file style errors. commit 459569a Author: ddavis-2015 <[email protected]> Date: Sun Oct 20 19:17:43 2024 -0700 use kernel optimzer level -O3 and -LNO:simd for Xtensa HIFI5 commit 122db20 Author: ddavis-2015 <[email protected]> Date: Sun Oct 20 19:07:19 2024 -0700 add compression build/test to bazel default test script commit d96b614 Author: ddavis-2015 <[email protected]> Date: Sun Oct 20 18:56:31 2024 -0700 fix CI code style errors fix CI BUILD file style errors commit 821dfdf Author: ddavis-2015 <[email protected]> Date: Sun Oct 20 12:19:34 2024 -0700 Cleanup header file usage. Add decompression code to Bazel BUILD files commit 4a02b22 Author: ddavis-2015 <[email protected]> Date: Sat Oct 19 13:31:11 2024 -0700 cleanup commit 3d765e6 Author: ddavis-2015 <[email protected]> Date: Fri Oct 18 18:01:21 2024 -0700 Squashed commit of the following: commit eaee851 Author: ddavis-2015 <[email protected]> Date: Fri Oct 18 17:48:48 2024 -0700 Squashed commit of the following: commit 4894265 Author: ddavis-2015 <[email protected]> Date: Fri Oct 18 17:48:05 2024 -0700 pre-merge empty commit commit a110e41 Author: ddavis-2015 <[email protected]> Date: Fri Oct 18 16:17:13 2024 -0700 fix C++ bitwidth 6 & 7 decompression commit efedcc2 Author: ddavis-2015 <[email protected]> Date: Fri Oct 18 10:18:50 2024 -0700 working decompression unit test commit 81ecf2e Author: ddavis-2015 <[email protected]> Date: Thu Oct 17 18:17:06 2024 -0700 decompression unit test improvements commit b318421 Author: ddavis-2015 <[email protected]> Date: Wed Oct 16 17:34:09 2024 -0700 add decompression unit test commit 9bb2b63 Author: ddavis-2015 <[email protected]> Date: Sun Oct 13 18:34:01 2024 -0700 cleanup commit 77bb05d Author: ddavis-2015 <[email protected]> Date: Sun Oct 13 18:29:33 2024 -0700 align compressed tensor data as per schema commit ad2b1c3 Author: ddavis-2015 <[email protected]> Date: Sat Oct 12 22:35:54 2024 -0700 reduce HIFI5 decompression code size commit 99c6e35 Author: ddavis-2015 <[email protected]> Date: Fri Oct 11 14:02:58 2024 -0700 revert to original Cadence bit width 4 code commit 2388549 Author: ddavis-2015 <[email protected]> Date: Thu Oct 10 17:50:29 2024 -0700 refactor decompression code into reference and platform specific Apply some Xtensa acceleration code changes commit b84853c Author: ddavis-2015 <[email protected]> Date: Tue Oct 8 16:08:55 2024 -0700 testing commit c107f42 Author: Ryan Kuester <[email protected]> Date: Thu Oct 17 14:31:03 2024 -0500 refactor: move misplaced TF_LITE_REMOVE_VIRTUAL_DELETEs to private: Move several TF_LITE_REMOVE_VIRTUAL_DELETE declarations that are wrongly in a public section of their classes. To have the intended effect, as documented in t/l/m/compatibility.h, these must be in a private section. commit 7b3a2bd Author: Ryan Kuester <[email protected]> Date: Thu Oct 17 12:36:46 2024 -0500 build(bazel): always build with TF_LITE_STATIC_MEMORY Add TF_LITE_STATIC_MEMORY to the defines set globally for TFLM builds in Bazel. TFLM always builds with this set in Make, and it appears to have been an oversight that it wasn't set during Bazel builds. Not having it set in Bazel caused some unit tests to pass under Bazel that failed under Make. At the same time, add -fno-exceptions. This flag is also always set in Make builds. Without it, setting TF_LITE_STATIC_MEMORY breaks the build. TF_LITE_STATIC_MEMORY triggers TF_LITE_REMOVE_VIRTUAL_DELETE in t/l/m/compatibility.h, which makes operator delete private in certain classes. When exceptions are enabled, a placement new with those classes is allowed to throw an exception, and operator delete is implicitly called during the unwind. The build breaks because operator delete can't be called if it's private. Disabling exceptions eliminates the unwind code that calls operator delete implicitly, and thus the build succeeds. In any case, -fno-exceptions should have been used in Bazel builds, matching the flags used in Make and the no-exceptions design requirement of the TFLM project. commit 1eb4e0d Author: Ryan Kuester <[email protected]> Date: Thu Oct 17 11:05:45 2024 -0500 feat(python): don't check .sparsity in interpreter Remove the check for sparse tensors in the Python interpreter wrapper. This fixes a broken build when TF_LITE_STATIC_MEMORY is set, which should always be the case in TFLM. TfLiteTensor objects don't have a .sparsity member when TF_LITE_STATIC_MEMORY is set. This prepares for an upcoming commit setting TF_LITE_STATIC_MEMORY during Bazel builds. This hasn't caused build failures in Make builds, which have always set TF_LITE_STATIC_MEMORY, because Make builds don't build the Python interpreter wrapper. commit 7217095 Author: Ryan Kuester <[email protected]> Date: Wed Oct 16 14:03:25 2024 -0500 fix(memory_arena_threshold): with TF_LITE_STATIC_MEMORY Fix the broken build due to redefinition of the threshold when TF_LITE_STATIC_MEMORY is set. Apparently this case isn't triggered in any Bazel test, only in Make. Simplify the threshold specification by only depending on whether compression is enabled and not also on whether TF_LITE_STATIC_MEMORY is in use. commit 8e4e55e Author: Ryan Kuester <[email protected]> Date: Thu Oct 10 12:38:03 2024 -0500 build(bazel): disable codegen when building --//:with_compression The codegen prototype code is not compatible with the changes which implement model compression made to the core TFLM components. For now, disable codegen targets when building with compression enabled. commit 884a234 Author: Ryan Kuester <[email protected]> Date: Tue Oct 15 18:31:01 2024 -0500 build(bazel): compile in compression when --//:with_compression Conditionally compile in support for compressed tensors when the option --//:with_compression is given. commit a1d459b Author: Ryan Kuester <[email protected]> Date: Thu Oct 10 12:28:39 2024 -0500 build(bazel): add --//with_compression build setting Add a --//with_compression user-defined build setting and a corresponding configuration setting. commit 4edc564 Author: Ryan Kuester <[email protected]> Date: Thu Oct 10 12:24:53 2024 -0500 build(bazel): fix compression-related dependencies of micro_allocator commit a52f97f Author: Ryan Kuester <[email protected]> Date: Tue Oct 15 17:28:09 2024 -0500 build(bazel): replace cc_* with tflm_cc_* in remaining TFLM code Replace cc_* targets remaining in TFLM code with tflm_cc_* targets. These are targets which did not formerly use the common copts. Avoid changing imported TFLite code, if for no other reason than to avoid merge conflicts during the automatic sync with upstream TFLite. commit a6368f4 Author: Ryan Kuester <[email protected]> Date: Fri Oct 11 16:08:34 2024 -0500 build(bazel): introduce tflm_cc_* macros, refactoring away micro_copts Remove micro_copts() by replacing every cc_* target that used them with a tflm_cc_* equivalent, and setting those common copts in one place, inside the tflm_cc_* macro. This is the first of several commits introducing tflm_cc_* macros in place of cc_binary, cc_library, and cc_test. Motivated by the upcoming need to support conditional compilation, the objective is to centralize build configuration rather than requiring (and remembering that) each cc_* target in the project add the same common attributes such as compiler options and select()ed #defines. Alternatives such as setting global options on the command line or in .bazelrc, even if simplified with a --config option, fail to preserve flags and hooks for configuration in the case TFLM is used as an external repository by an application project. Nor is it easy in that case for individual targets to override an otherwise global setting. commit 1518422 Author: Ryan Kuester <[email protected]> Date: Thu Oct 10 23:56:49 2024 -0500 chore: remove obsolete ci/temp_patches Remove ci/temp_patches, which was obsoleted in 23f608f once it was no longer used by the sync script. It should have been deleted then. Remove it not only to clean up dead code, but because it contains a reference to `micro_copts`, which is about to be refactored away, and we don't want to leave stray references to it in the tree. commit 18ef080 Author: Ryan Kuester <[email protected]> Date: Tue Oct 8 17:58:12 2024 -0500 refactor: use metadata_saved.h instead of metadata_generated.h Use the generated file metadata_saved.h instead of metadata_generated.h for the reasons explained in t/l/m/compression/BUILD:metadata_saved. Delete metadata_generated.h from the source tree as it is not maintained. commit 5a02e30 Author: Ryan Kuester <[email protected]> Date: Thu Oct 10 13:46:46 2024 -0500 test(memory_arena_threshold): adjust expected value with compression Fix a test failure by setting a different expected value for the persistent buffer allocation when compression is configured in. The allocation was allowed to vary by 3%; however, compression adds ~10%. Set the expected value to the measured value when compression is configured in. commit 01bc582 Author: Ryan Kuester <[email protected]> Date: Thu Oct 10 13:35:10 2024 -0500 test(memory_arena_threshold): don't expect exact allocation values Remove the check for allocation sizes to exactly match expected values. This check immediately followed--and thus rendered pointless---a check that sizes are within a certain percentage, which seems to be the true intent of the test. commit e0aae77 Merge: e328029 e86d97b Author: Ryan Kuester <[email protected]> Date: Wed Oct 16 13:39:56 2024 -0500 Merge branch 'main' into compress-testing commit e328029 Author: Ryan Kuester <[email protected]> Date: Mon Oct 7 12:52:23 2024 -0500 build(bazel): fix dependencies in work-in-progress compression code In the Bazel build, add dependencies needed by the code added to t/l/m:micro_context for decompression. The Bazel build with or without compression was broken without this. commit e86d97b Author: RJ Ascani <[email protected]> Date: Mon Oct 7 10:36:26 2024 -0700 Replace rascani with suleshahid on OWNERS (tensorflow#2715) BUG=none commit b773428 Author: Ryan Kuester <[email protected]> Date: Fri Oct 4 09:59:10 2024 -0500 feat(compression): add work-in-progress compression and viewer tools commit f6bd486 Merge: 487c17a e3f6dc1 Author: Ryan Kuester <[email protected]> Date: Fri Oct 4 09:36:24 2024 -0500 Merge branch 'main' into compress-prerelease commit e3f6dc1 Author: David Davis <[email protected]> Date: Thu Oct 3 10:45:00 2024 -0700 Compression documentation (tensorflow#2711) @tensorflow/micro Add documentation describing some compression/decompression internals and makefile build procedures. bug=tensorflow#2710 commit b3967a9 Author: Ryan Kuester <[email protected]> Date: Wed Oct 2 13:36:01 2024 -0500 style: add .style.yapf to control yapf styling of Python code (tensorflow#2709) Add a .style.yapf file so yapf can be used to style Python code without passing the project's style via command line option. Remove the corresponding patch to pigweed's call to yapf, used by CI, and instead let it too rely on .style.yapf. Remove the developer documentation's instruction to use the command line option. BUG=description commit d249577 Author: Ryan Kuester <[email protected]> Date: Tue Oct 1 16:16:45 2024 -0500 build(codegen): suppress noise in console output (tensorflow#2708) Add a --quiet option to the code_generator binary so that when it's used within the build system, it doesn't print unexpected, distracting noise to the console. Generally, compiler or generator commands don't print output unless there's an error. BUG=description commit 4894265 Author: ddavis-2015 <[email protected]> Date: Fri Oct 18 17:48:05 2024 -0700 pre-merge empty commit commit a110e41 Author: ddavis-2015 <[email protected]> Date: Fri Oct 18 16:17:13 2024 -0700 fix C++ bitwidth 6 & 7 decompression commit efedcc2 Author: ddavis-2015 <[email protected]> Date: Fri Oct 18 10:18:50 2024 -0700 working decompression unit test commit 81ecf2e Author: ddavis-2015 <[email protected]> Date: Thu Oct 17 18:17:06 2024 -0700 decompression unit test improvements commit b318421 Author: ddavis-2015 <[email protected]> Date: Wed Oct 16 17:34:09 2024 -0700 add decompression unit test commit 9bb2b63 Author: ddavis-2015 <[email protected]> Date: Sun Oct 13 18:34:01 2024 -0700 cleanup commit 77bb05d Author: ddavis-2015 <[email protected]> Date: Sun Oct 13 18:29:33 2024 -0700 align compressed tensor data as per schema commit ad2b1c3 Author: ddavis-2015 <[email protected]> Date: Sat Oct 12 22:35:54 2024 -0700 reduce HIFI5 decompression code size commit 99c6e35 Author: ddavis-2015 <[email protected]> Date: Fri Oct 11 14:02:58 2024 -0700 revert to original Cadence bit width 4 code commit 2388549 Author: ddavis-2015 <[email protected]> Date: Thu Oct 10 17:50:29 2024 -0700 refactor decompression code into reference and platform specific Apply some Xtensa acceleration code changes commit b84853c Author: ddavis-2015 <[email protected]> Date: Tue Oct 8 16:08:55 2024 -0700 testing
ddavis-2015
added a commit
to ddavis-2015/tflite-micro
that referenced
this issue
Nov 25, 2024
commit 414a55a Author: ddavis-2015 <[email protected]> Date: Thu Nov 21 18:15:40 2024 -0800 last merge missed xtensa.inc file commit 1110543 Author: ddavis-2015 <[email protected]> Date: Thu Nov 21 18:06:45 2024 -0800 Squashed commit of the following: commit 300751d Author: ddavis-2015 <[email protected]> Date: Wed Nov 20 12:03:16 2024 -0800 Update to latest Cadence code. Int8 any bitwidth on normal quant axis updated. commit 0a49b2a Author: ddavis-2015 <[email protected]> Date: Wed Nov 20 10:50:30 2024 -0800 Add input tensor CRC to Generic Benchmark application. Only use -O3 -LNO:simd with the xtensa decompress.cc target. commit 83dafce Author: ddavis-2015 <[email protected]> Date: Mon Nov 18 17:23:58 2024 -0800 cleanup commit 2f8cead Author: ddavis-2015 <[email protected]> Date: Mon Nov 18 13:29:33 2024 -0800 Revert FakeMicroContext changes for alternate profiler. Add default alternate profiler implementation to MicroContext. commit fddf003 Author: ddavis-2015 <[email protected]> Date: Mon Nov 18 12:29:26 2024 -0800 Fix typo. commit ae6a207 Author: ddavis-2015 <[email protected]> Date: Mon Nov 18 12:29:06 2024 -0800 Implement alternate profiler for MicroInterpreter. Enable use of alternate profiler by decompression code. Enable use of alternate profiler by Generic Benchmark application. commit 5e1a1c9 Author: ddavis-2015 <[email protected]> Date: Sun Nov 10 18:24:02 2024 -0800 changes to make the memory planner debug output easier to interpret commit f651c88 Author: ddavis-2015 <[email protected]> Date: Sun Nov 10 04:27:29 2024 -0800 single pending ops queue process pending ops recursively commit cfd9890 Author: ddavis-2015 <[email protected]> Date: Sun Nov 10 00:26:14 2024 -0800 expand model_facade redo var handle tracking commit 7776cda Author: ddavis-2015 <[email protected]> Date: Tue Nov 5 13:24:58 2024 -0800 remove [[maybe_unused]] commit 40e7530 Author: ddavis-2015 <[email protected]> Date: Mon Nov 4 11:17:43 2024 -0800 fix arena commit 0d889e0 Author: ddavis-2015 <[email protected]> Date: Sat Nov 2 14:18:55 2024 -0700 Fix MicroProfiler bug with ClearEvents(). Add pre-inference profiling to the Generic Benchmark. commit bd04eb2 Author: ddavis-2015 <[email protected]> Date: Wed Nov 20 18:40:42 2024 -0800 Pre-rebase empty commit commit 661ad5e Author: Ryan Kuester <[email protected]> Date: Thu Nov 14 11:31:23 2024 -0600 feat(compression): update model compression tools commit 6eac102 Author: Ryan Kuester <[email protected]> Date: Thu Nov 14 00:31:56 2024 -0600 feat: add model compression draft Add a feature-complete draft of the model compression feature in a single commit. This commit isn't intended for merging into mainline; it exists solely for reviewers and testers to take an early look. commit f91dd91 Author: Ryan Kuester <[email protected]> Date: Wed Nov 13 16:46:20 2024 -0600 build(codegen): suppress noise in console output (tensorflow#2742) build(codegen): suppress noise in console output Redirect output from the code_generator binary so that when it's used within the build system, it doesn't print unexpected, distracting noise to the console. Generally, compiler or generator commands don't print output unless there's an error. This reverts "build(codegen): suppress noise in console output (tensorflow#2708)", commit d249577, which attempted to fix the same problem by adding a --quiet flag. A subsequent upgrade to the Tensorflow Python package caused new noise that wasn't possible to suppress with the flag. BUG=description commit 182c8c7 Author: Adrian Lundell <[email protected]> Date: Tue Nov 12 23:34:09 2024 +0100 Update CMSIS-NN Transpose conv function call (tensorflow#2760) Also removes OPTIMIZE_FOR_SIZE compiler option as the new implementation is more space efficient. BUG=CMSIS-NN update commit d7c42b9 Author: suleshahid <[email protected]> Date: Tue Nov 12 12:40:15 2024 -0800 Add SQRT to Requantize Flatbuffer. (tensorflow#2756) BUG=Add more support for tool. commit 26ada36 Author: Jae H. Yoo <[email protected]> Date: Tue Nov 12 12:14:54 2024 -0800 Update numpy_utils.cc to support ml_dtypes.bfloat16 (tensorflow#2758) This commit adds NPY_USERDEF for ml_dtypes.bfloat16 only. Other types are not supported yet. BUG=tensorflow#2759 BUG=2759 commit 79c3fde Author: Ryan Kuester <[email protected]> Date: Tue Nov 12 13:40:59 2024 -0600 ci(bazel): improve logging of bazel CI tests (tensorflow#2749) Improve the logging of Bazel commands run in CI by setting several options that turn off, e.g., progress messages intended for interactive shells, and turn on, e.g., more verbose output when something fails. BUG=tensorflow#2742 commit 9245002 Author: Ryan OShea <[email protected]> Date: Wed Nov 6 20:18:28 2024 +0100 Fix issue with transposing shape in CMSIS-NN batch matmul (tensorflow#2741) BUG=tensorflow#2740 commit 4bb78c7 Author: Ryan Kuester <[email protected]> Date: Tue Nov 5 19:11:45 2024 -0600 fix(create_tflm_tree): remove recent tests from exported tree (tensorflow#2751) Remove the recently added span_test.cc and static_vector_test.cc from the files exported by the create_tflm_tree.py project generation process by adding them to the list of tests in the Makefile. Unit tests are not meant to be included in exported trees; they may include files that are not exported. This change also ensures that these tests are included when `make test` is run. BUG=fixes tensorflow#2718 commit 45cd79b Author: Måns Nilsson <[email protected]> Date: Tue Nov 5 23:13:20 2024 +0100 Replace CoreDebug with DCB (tensorflow#2746) BUG=The CMSIS CoreDebug macro is deprecated. commit 694d250 Author: Ryan Kuester <[email protected]> Date: Tue Nov 5 15:48:52 2024 -0600 chore: remove obsolete ci/temp_patches (tensorflow#2744) chore: remove obsolete ci/temp_patches Remove ci/temp_patches, which was obsoleted in 23f608f once it was no longer used by the sync script. It should have been deleted then. Remove it not only to clean up dead code, but because it contains a reference to `micro_copts`, which is about to be refactored away, and we don't want to leave stray references to it in the tree. BUG=tensorflow#2636 commit 8eb6b23 Author: Ryan Kuester <[email protected]> Date: Fri Nov 1 17:10:05 2024 -0500 build(bazel): add integrity check to nnlib_hifi4 download (tensorflow#2743) build(bazel): add integrity check to nnlib_hifi4 download Add an integrity check to the http_archive() download of nnlib_hifi4 in order to make the build more hermetic, reduce the security risk that a remote file changes, and silence the noisy warning on the console during the build: DEBUG: Rule 'nnlib_hifi4' indicated that a canonical reproducible form can be obtained by modifying arguments integrity[....] BUG=description commit 5a46ab0 Author: Måns Nilsson <[email protected]> Date: Fri Nov 1 20:57:00 2024 +0100 Add kernels optimized for size flag to FC and SVDF (tensorflow#2734) The kernels optimized for size flag provides an alternative implementation where size is prioritized over latency. For size option (speed option is default) it means the CMSIS-NN kernels are calculating kernel sums during inference. BUG=no bug but this will let users prioritize speed vs size even more commit 389e775 Author: suleshahid <[email protected]> Date: Tue Oct 29 13:57:29 2024 -0700 Fix upstream sync (tensorflow#2728) * . * nl add * Sync from upstream TF. * fix include --------- Co-authored-by: TFLM-bot <[email protected]> commit e440f0a Author: Måns Nilsson <[email protected]> Date: Mon Oct 28 20:16:43 2024 +0100 Update CMSIS third party download (tensorflow#2716) CMSIS is updated from CMSIS_5 to CMSIS_6. BUG=It is recommended to upgrade from CMSIS_5 to CMSIS_6 commit 1b975d9 Author: suleshahid <[email protected]> Date: Mon Oct 28 11:27:23 2024 -0700 Update deprecated mergify rules (tensorflow#2732) * . * nl add * update mergify commit 78e5467 Author: Ryan Kuester <[email protected]> Date: Fri Oct 25 18:27:21 2024 -0500 fix(python): add compatibility with NumPy 2 (tensorflow#2733) Upgrade the pinned packages used in the Bazel build environment to make the Python extension module compatible with NumPy 2 in addition to NumPy 1. Modules built against NumPy 1 are not compatible with runtime environments that use NumPy 2, but modules built against NumPy 2 should be compatible with both. The python_requirements.txt defining the Bazel build environment is was upgraded via the command: bazel run //third_party:python_requirements.update -- --upgrade See third_party/python_requirements.in for details. This change upgraded a number of other packages as well, notably Tensorflow, from 2.17 to 2.18. In fact, Tensorflow 2.18 is also required for compatibility with NumPy 2. Also update the path to the header files within the NumPy wheel, used by the :numpy_cc_deps target. Our method of including header files shipped in Python packages is sensitive to the internal details of those packages, and NumPy moved their headers during this major version change. This fixes tensorflow#2731, in which //python/tflite_micro:whl_test began failing. :whl_test installs the newly built tflite_micro package, which packages our Python extension module, and its dependencies into a clean virtual environment and performs tests. The tflite_micro package has unversioned dependencies on the tensorflow and numpy packages. Within the last 24 hours, tensorflow in PyPI was upgraded from 2.17 to 2.18. The 2.18 release upgraded tensorflow's versioned dependency on numpy from 1 to 2, forcing the environment created by :whl_test to use NumPy 2 instead of NumPy 1; thereby exposing the incompatibility of the extension module that was built against NumPy 1, and causing :whl_test to fail. BUG=tensorflow#2731 commit e86d97b Author: RJ Ascani <[email protected]> Date: Mon Oct 7 10:36:26 2024 -0700 Replace rascani with suleshahid on OWNERS (tensorflow#2715) BUG=none commit e3f6dc1 Author: David Davis <[email protected]> Date: Thu Oct 3 10:45:00 2024 -0700 Compression documentation (tensorflow#2711) @tensorflow/micro Add documentation describing some compression/decompression internals and makefile build procedures. bug=tensorflow#2710 commit b3967a9 Author: Ryan Kuester <[email protected]> Date: Wed Oct 2 13:36:01 2024 -0500 style: add .style.yapf to control yapf styling of Python code (tensorflow#2709) Add a .style.yapf file so yapf can be used to style Python code without passing the project's style via command line option. Remove the corresponding patch to pigweed's call to yapf, used by CI, and instead let it too rely on .style.yapf. Remove the developer documentation's instruction to use the command line option. BUG=description commit d249577 Author: Ryan Kuester <[email protected]> Date: Tue Oct 1 16:16:45 2024 -0500 build(codegen): suppress noise in console output (tensorflow#2708) Add a --quiet option to the code_generator binary so that when it's used within the build system, it doesn't print unexpected, distracting noise to the console. Generally, compiler or generator commands don't print output unless there's an error. BUG=description
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
@tensorflow/micro
Add documentation describing some compression/decompression internals and makefile build procedures.
The text was updated successfully, but these errors were encountered: