Age | Commit message (Collapse) | Author |
|
- Update TensorFlow dependencies for x86_64
- Adapt unit tests to new TensorFlow version
- Update linters (including pre-commit hooks) and fix issues
- Use conditional import to fix tflite compat code for aarch64
Change-Id: I1a9b080b900ab65e38f7f2552562822bbfdcd259
|
|
Check input model for Arm NN TensorFlow Lite Delegate 22.08 support.
Change-Id: I1253c4c0b294c5283e08f0a39561b922ef0f62e6
|
|
Change-Id: I212c9cad5f2ac28b75b1d9e2fb8f1f8b444b8397
|
|
* Add `api.generate_supported_operators_report` that calls
a report generator based on given target profile.
* Update `--supported-ops-report` cli flag help section and README.md
info by specifying that it is currently only supported for Ethos-U.
* Add tests for `generate_supported_operators_report` method.
* Add stub methods `report` for Cortex-A
and Tosa operators.py module that throw a "Not supported" exception.
Change-Id: Iaf373fe48299d807e48fd6a0e6c029ce32203f4f
|
|
- Add ability to intercept low level TensorFlow output
- Produce advice for the models that could not be
converted to the TensorFlow Lite format
- Refactor utility functions for TensorFlow Lite
conversion
- Add TensorFlow Lite compatibility checker
Change-Id: I47d120d2619ced7b143bc92c5184515b81c0220d
|
|
Due to the new version of Corstone-310, the generic inference runner
has been updated to 22.08:
* The inference runner 22.08 doesn't have timing adapters because
incompatible with new version of Corstone-310
* Remove some memory mode logic which is not needed anymore
* Corstone-310 allows to run Ethos-U65 simulation hence adding the
binary for enabling this case
* Delete the inference runner 22.05
* Fix codebase and tests to cope with changes above
Change-Id: I3dc894d7cb49b09102a19b0d7f588694a0f3b99f
|
|
* Add Cortex-A device skeleton
* Add unit tests for the Cortex-A device skeleton
* Update profiles.json by adding the new "cortex-a" profile
* Add new cortex-a factory to the get_advisor method in api.py
* Disable performance and optimization commands for the cortex-a
profile.
* Update trademarks section in README.md
* Update pyproject.toml to not run similarity check in imports
Change-Id: I2e228aaada1e2d3c5cc329d70572b51962ff517f
|
|
Use "TensorFlow Lite" instead of "TFLite" in
documentation and comments
Change-Id: Ie4450d72fb2e5261d152d72ab8bd94c3da914c46
|
|
Change-Id: I2ce8ccebf8a9972d9936164ae52912044c682633
|
|
- Update configuration for inclusive language linter
- Fix reported issues
Change-Id: If0f8b6e20c17d8ee1c6179c61040fc351437f036
|
|
- Remove the extra 's' character which broke the path detection
mechanism when installed from /opt/VHT location on AVH instances
- Add unit tests to make sure we fixed the problem
Change-Id: Ic33a3c6c4d2b33181268d0efbc0324962e741dd4
|
|
- added the target profile for ethos-u65-256
- modified the unit tests regarding target profiles
- updated the README.md
Change-Id: I0f6b50199e735a18f0b436d291c871e1fa0460dd
|
|
- Use pyupgrade tool for the syntax upgrade
Change-Id: I3b9ae6cc5da193329e37028cde967ae049388361
|
|
- Enable deferred annotations evaluation
- Use builtin types for type hints whenever possible
- Use | syntax for union types
- Rename mlia.core._typing into mlia.core.typing
Change-Id: I3f6ffc02fa069c589bdd9e8bddbccd504285427a
|
|
Apply the same formatting rules for each section in the
output of the command "backend status"
Change-Id: I253c513f99826ef1eb916285ca4c2cc1840633ac
|
|
- mypy issue: to make the comment #type: ignore platform specific,
flags like platform.machine() cannot be recognized by mypy, so we
cannot isolate the specific lines of code that fail mypy tests
- numpy issue: for numpy version < 1.20, the function np.unique
has not been type annotated, which caused mypy throwing the error
when we run our unit tests in aarch64
- because of the above two reasons, we use function decorator to
turn off type checking for entire functions to remove all
annotations so that the mypy error for certain lines can be silented
Change-Id: Id91e65ef7677b78b4c9c85b8412229e3672e3a66
|
|
Change-Id: I9dfa9f17a93ac24be0aa9ad68e4a72b79aecfb9d
|
|
- Add new module for TOSA
- Add advisor workflow components
- Use TOSA checker for getting operators compatibility
information
Change-Id: I769e5e2a84e15779658f0895b4a347384def63bf
|
|
- Move target specific details out of API module
- Move common logic for workflow event handler into a
separate class
Change-Id: Ic4a22657b722af1c1fead1d478f606ac57325788
|
|
Add bandit to pre-commit and fix some bandit errors.
We use the default security level (low) with few exceptions:
* B101 assert_use: apart of tests, we use assert in our codebase
hence we globally ignore error B101.
* B404/B603: these are errors related to subprocesse and they are
being ignored locally when used.
* B604 Test for any function with shell equals true: we have disabled
this locally because of its safe use in the tests.
Change-Id: If654e5e92285f7c86ac210a6f1373dbab6be17c9
|
|
Re-factoring the code base to further merge the old AIET code into MLIA.
- Remove last traces of the backend type 'tool'
- Controlled systems removed, including SSH protocol, controller,
RunningCommand, locks etc.
- Build command / build dir and deploy functionality removed from
Applications and Systems
- Moving working_dir()
- Replace module 'output_parser' with new module 'output_consumer' and
merge Base64 parsing into it
- Change the output consumption to optionally remove (i.e. actually
consume) lines
- Use Base64 parsing in GenericInferenceOutputParser, replacing the
regex-based parsing and remove the now unused regex parsing
- Remove AIET reporting
- Pre-install applications by moving them to src/mlia/resources/backends
- Rename aiet-config.json to backend-config.json
- Move tests from tests/mlia/ to tests/
- Adapt unit tests to code changes
- Dependencies removed: paramiko, filelock, psutil
- Fix bug in corstone.py: The wrong resource directory was used which
broke the functionality to download backends.
- Use f-string formatting.
- Use logging instead of print.
Change-Id: I768bc3bb6b2eda57d219ad01be4a8e0a74167d76
|
|
- Update TensorFlow to version 2.8 (now supported by Vela 3.4)
- Adapt existing codebase to preserve intermediary tensors in the interpreter in order to avoid errors when trying to print all of them in the future.
- Ignore types for numpy methods that do not have typing annotations in their definitions. This is needed because otherwise mypy would consider the calling function to also be untyped.
Change-Id: I943ac196fd4e378f5238949b15c23a2d628c8b5e
|
|
Upgrade Vela version from 3.3.0 to 3.4.0.
- Adapt code to use new typing notation by replacing `numpy.array` with `numpy.ndarray` where necessary.
Change-Id: I035e9564d448652aa09a52d79c71ef09663ea776
|
|
Merge the deprecated AIET interface for backend execution into MLIA:
- Execute backends directly (without subprocess and the aiet CLI)
- Fix issues with the unit tests
- Remove src/aiet and tests/aiet
- Re-factor code to replace 'aiet' with 'backend'
- Adapt and improve unit tests after re-factoring
- Remove dependencies that are not needed anymore (click and cloup)
Change-Id: I450734c6a3f705ba9afde41862b29e797e511f7c
|
|
When running the e2e test of Wav2Letter, the tensor allocation failed.
To resolve this, we changed the arena cache size to 2096768. This is
due to the requirement in Vela. We need to make it aware of the memory
limitation on the target we are deploying for.
Change-Id: I192790033954058fa5db249452d998764df52663
|
|
Fix the issue that no performance information is shown for TFLite files
when the mode 'all_tests' is used.
Change-Id: I8b4df4ab84ba9783b582ad449a34bf6177037e14
|
|
Use TFLITE_BUILTINS instead of TFLITE_BUILTINS_INT8 to make the
quantization non-strict.
Note: De facto this does not change the behavior of MLIA because the
TFLITE_BUILTINS_INT8 is not correctly applied with the new quantizer.
See: https://github.com/tensorflow/tensorflow/issues/56535
Change-Id: Ia0782ba22c5e9223fa10fec71c16aee60b79bb57
|
|
Add MLIA codebase including sources and tests.
Change-Id: Id41707559bd721edd114793618d12ccd188d8dbd
|