- 28 8月, 2021 1 次提交
-
-
由 Nat Jeffries 提交于
-
- 27 8月, 2021 2 次提交
-
-
由 Nat Jeffries 提交于
-
由 Advait Jain 提交于
There isn't any need to have a special define for the target. In fact, having such a define can create trouble when the TARGET and OPTIMIZED_KERNEL_DIR have the same name. Priot to this change, -D<target> and -D<optimized_kernel_dir> could result in the same define which would interfere with the kernel variant definitions (which are tied to the optimized_kernel_dir and are independent of the target). Manually tested that the following command fails without this change, and passes with it: ``` make -f tensorflow/lite/micro/tools/make/Makefile TARGET=xtensa TARGET_ARCH=hifi4 XTENSA_CORE=F1_190305_swupgrade -j8 keyword_benchmark ``` BUG=http://b/197889069
-
- 26 8月, 2021 2 次提交
-
-
由 deqiangc 提交于
* Make size check workflow error out if memory footprint increase between head of main and this PR. This error can be disabled by putting MEM_CHANGE_OK= tag. In either case, a memory footprint comparison report will be produced as part of this workflow action. The plan is for the code reviewer to audit the report when the PR increase memory footprint and only allow the PR that increase memory footprint to be merged with MEM_CHANGE_OK=what_is_the_reason_for_this_increase_and_why_this_is_ok The PR shall be up to date for the comparison to be accurate. So before applying MEM_CHANGE_OK=, please make the PR branch up to date. BUG=http://b/196637015 * Fix an if condition
-
由 dependabot[bot] 提交于
Bumps [tensorflow](https://github.com/tensorflow/tensorflow) from 2.5.0 to 2.5.1. - [Release notes](https://github.com/tensorflow/tensorflow/releases) - [Changelog](https://github.com/tensorflow/tensorflow/blob/master/RELEASE.md) - [Commits](https://github.com/tensorflow/tensorflow/compare/v2.5.0...v2.5.1) --- updated-dependencies: - dependency-name: tensorflow dependency-type: direct:production ... Signed-off-by: Ndependabot[bot] <support@github.com> Co-authored-by: Ndependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
-
- 25 8月, 2021 2 次提交
-
-
由 Måns Nilsson 提交于
* Updating downloaded CMSIS version. * Removing -Wno-strict-aliasing for CMSIS-NN. Change-Id: I41a99fd8d5aa3bf5167a42adfc1ae02449a071a0 Co-authored-by: NAdvait Jain <advaitjain@users.noreply.github.com>
-
由 Advait Jain 提交于
This reverts commit 1ca28598, reversing changes made to b703c9f9. Revert created with the following command: ``` git revert 1ca28598 -m 2 ```
-
- 24 8月, 2021 3 次提交
-
-
由 Chris Knorowski 提交于
Co-authored-by: NChris Knorowski <chris.knorowski@sensiml.com>
-
由 Måns Nilsson 提交于
* Replace std::partial_sort with merge_sort in PPD TFL has changed to std::stable_sort but TFLM can not use it because it is using the heap. Instead TFLM use merge sort. Change-Id: I015ee890214b65ebb14cac3b1e888b8f4213ebfb * Add unit test for stable sort in detection_postprocess Co-authored-by: NAdvait Jain <advaitjain@users.noreply.github.com>
-
由 Jakub Jatczak 提交于
-
- 20 8月, 2021 7 次提交
-
-
由 Advait Jain 提交于
With https://github.com/tensorflow/tflite-micro/pull/337 we can now run these targets via a bazel build. BUG=http://b/186676432 and http://b/174680668
-
由 Chris Knorowski 提交于
Co-authored-by: NChris Knorowski <chris.knorowski@sensiml.com>
-
由 Nat Jeffries 提交于
-
由 Nat Jeffries 提交于
-
由 Advait Jain 提交于
And restore -layering_check to a few places where it got removed with https://github.com/tensorflow/tflite-micro/pull/435/ BUG=http://b/177257332
-
由 Nat Jeffries 提交于
-
由 Nat Jeffries 提交于
* Fix misc formatting and spelling errors. * Re-enable layering check. Co-authored-by: NAdvait Jain <advaitjain@users.noreply.github.com>
-
- 18 8月, 2021 7 次提交
-
-
由 cad-audio 提交于
Currently script uses /tmp directory for the the xtensa downloads. Modifying sciprt to resolve the issue, mktemp creates unique folder. BUG=usage of /tmp directory for the xtensa downloads.
-
由 Advait Jain 提交于
* Build all the benchmarks as part of a `make build` command. * Build benchmark as part of `make build`. BUG=http://b/197017812 * Fix the build * Add a separate `make build` to the CI scripts.
-
由 Nat Jeffries 提交于
-
由 deqiangc 提交于
* Update size_check.yml and rename fusion_f1.json to more general binary_size.json. Updte the size_check workflow to be able to launched it manually for test. Also update it so that it can build image. The next step is to test whether this will lead to two PR to have merge conflict always. BUG=http://b/196637015 * Remove currently unused code
-
由 deqiangc 提交于
* Create a test used for binary size test. Also create a script to generate size report of a such binary. Modify the size_comp.py to be more generic (not limited to xtensa). what is tensorflow/lite/micro/tools/ci_build/binary_size_test will be eventually removed once we verify the feature. The next PR will update the yml file for the size_check.yml. BUG=http://b/196637015 * Restore lines that are deleted by mistakes * Update comments
-
由 Nat Jeffries 提交于
-
由 Nat Jeffries 提交于
-
- 17 8月, 2021 6 次提交
-
-
由 Advait Jain 提交于
BUG=cleanup
-
由 deqiangc 提交于
* Enable micro_interpreter_test and two other tests for stm32f4 TESTED=make -j8 -f tensorflow/lite/micro/tools/make/Makefile TARGET= stm32f4 test BUG=http://b/174707181 * Update todo BUG=http://b/158651472
-
由 cad-audio 提交于
* REF_CODE_REFACTOR: Leaky_Relu refactoring the reference code for leaky_relu operator. BUG=refactoring existing code. * small nits. * fix the formatting. Co-authored-by: NAdvait Jain <advaitjain@google.com>
-
由 cad-audio 提交于
* REF_CODE_REFACTOR: Circular_buffer refactoring the refernce code for circular_buffer operator. BUG=refactoring existing code. * Formatted the code. Co-authored-by: NAdvait Jain <advaitjain@users.noreply.github.com> Co-authored-by: NAdvait Jain <advaitjain@google.com>
-
由 Advait Jain 提交于
BUG=Fixes #418
-
由 Måns Nilsson 提交于
Change path to person detect model header file from example directory to where it is downloaded.
-
- 14 8月, 2021 7 次提交
-
-
由 Advait Jain 提交于
* Some clean up of the test script. * Have make fail if python script has an error. Prior to this change, the generate_cc_arrays.py script could fail (for example because PIL not being installed) but the error would be silently ignored. Instead, the build would fail much later because of missing header files and the root cause would be hard to debug. Follow-up PR will also use this error handling approach for other places where we call into Python and bash scripts from the make. BUG=http://b/196630709 * support earlier versions of make.
-
由 Advait Jain 提交于
* Have make also fail if Python script fails. BUG=http://b/196630709 * debugging error. * support versions of make < 4.2 * replace some tabs with spaces.
-
由 Meghna Natraj 提交于
* Update the download script for TFLite Micro person detection example * Update the download script for TFLite Micro person detection example
-
由 Advait Jain 提交于
-
由 Advait Jain 提交于
-
由 Advait Jain 提交于
* Add option to only do a build for the vision_p6. BUG=http://b/196459172 * Only build for Vision P6 as part of presubmit checks. * make running the tests optional. * fix whitespace. * Address review comments.
-
由 Advait Jain 提交于
BUG=Fixes #401
-
- 13 8月, 2021 3 次提交
-
-
由 Nat Jeffries 提交于
-
由 deqiangc 提交于
Create patch to flatbuffers that enables micro_interpreter_test and memory_helpers_test to run for bluepill architecture (#398) * Patch flatbuffers to avoid the use of new/delete. Also patch test_helpers to avoid the virtual delete. This enables micro_interpreter_test to run for bluepill platform. BUG=http://b/174707181 * Create patch that enable micro_interpreter_test to run. The patch assumes all TFLM code provides allocator for allocate buffer. * Update patch without extra formatting
-
由 deqiangc 提交于
Add comments to some magic number in micro_interpreter_test and also attempt to relax some checks (#394) * Add comments to some magic number in micro_interpreter_test * Relax some check on the micro_interpreter_test to enable it pass on xtensa and also add comments on how magic number comes from. BUG=http://b/174707181 * Remove redundant header files * Change from _GE to _LE for better readability
-