Skip to content
  1. Oct 11, 2023
  3. Oct 10, 2023
  5. Oct 05, 2023
  7. Oct 03, 2023
    • Johan Alfvén's avatar
      MLBEDSW-8102: Fix regression on Argmax int64 · 7972ee80
      Johan Alfvén authored 
      

- Fixed a regression where DepthWiseConv used in argmax int64
had the wrong shape.

- The error was introduced when adding support for a new operator
that changed the weight shape for the cast utility function. That
change only worked because reorder_depthwise_weights was called
later. Since argmax is converted after reorder_depthwise_weights
the cast operator in argmax got the wrong shape.

- The fix is to set the correct weight shape in the cast operator
and then mark that the weights already have been transposed correctly.

Change-Id: I61f5694f078cfcaf0d46d43faead6eb7e0a23ade
Signed-off-by: Johan Alfvén's avatarJohan Alfven <johan.alfven@arm.com>
      7972ee80
  9. Sep 18, 2023
  11. Sep 14, 2023
  13. Sep 13, 2023
  15. Sep 12, 2023
  17. Sep 06, 2023
  19. Sep 05, 2023
  21. Aug 29, 2023
  23. Aug 22, 2023
    • Johan Alfvén's avatar
      MLBEDSW-7949: [MLCE] Remove duplicate cpu tensors · c02eaa3e
      Johan Alfvén authored 
      

- If a npu op is followed by a convolution op than runs on the cpu,
the optimized file ends up containing a duplicated tensor called _cpu.
Functionality wise not a problem but the graph will look strange in a
graph viewer.

- This error was introduced when removing duplicate weights
tensors but the above use case was not considered in that patch.

- The fix is to make sure that only the weight and bias tensor are
modified.

Change-Id: I576f13650f1f9d3d50a421ab7100fc8b5ab62657
Signed-off-by: Johan Alfvén's avatarJohan Alfven <johan.alfven@arm.com>
      c02eaa3e
  25. Aug 21, 2023
    • Rob Elliott's avatar
      Moving Vela to use TOSA v0.80.0 specification · 00a15db3
      Rob Elliott authored and Rickard Bolin's avatar Rickard Bolin committed 
      

 * Using serialization_lib main branch to update statically copied
   files sha 5f920211ac23393a7b98a0d358bfbfc3232d5c8f (v0.80.0)
 * All files within the ethosu/vela/tosa are copied from that revision
 * Note: hope to move to serialization_lib as a pip module in future

 * Modified the ethosu/vela/{tosa_mapping,tosa_reader}.py to use
   v0.80.0 TOSA FlatBuffers implementation
 * These are the additional changes made to support this new version,
   with changes in the format of the FlatBuffers file and where various
   values are stored. Either changing from input to attribute, or
   moving to different attributes.

Signed-off-by: Rob Elliott's avatarRob Elliott <robert.elliott@arm.com>
Change-Id: I5e1fcc2a9964148619be3477adf1e88e84cbae2d
      00a15db3
    • Rickard Bolin's avatar
      MLBEDSW-7702: Update release notes · 8ea90edb
      Rickard Bolin authored 
      

- Added release information
- Modified SUPPORTED_OPS.md version info
- Update README.md and classifiers in pyproject.toml to specify Python
  3.10 as recommended and tested version

Change-Id: I78e5752846f261d4713b89c8efe447bcb9c095dd
Signed-off-by: Rickard Bolin's avatarRickard Bolin <rickard.bolin@arm.com>
      8ea90edb
  27. Aug 16, 2023
    • Johan Alfvén's avatar
      MLBEDSW-7884: Fix crash for RSQRT · 3db30ff5
      Johan Alfvén authored 
      

- RSQRT is only defined for positive numbers and
therefore the zeropoint and actual input value
will have an impact

- Clamp the range to avoid crashing. As long as the actual
input is within valid range everything works. If the input
is not valid the reference will crash and not generating
any output

Change-Id: I1082b508d9cd85ad4b017e7b786cfff730585172
Signed-off-by: Johan Alfvén's avatarJohan Alfven <johan.alfven@arm.com>
      3db30ff5
  29. Aug 10, 2023
  31. Aug 09, 2023
  33. Aug 08, 2023
    • Tim Hall's avatar
      MLBEDSW-7689: Document verbose command stream options · cd03504c
      Tim Hall authored 
      

 - Documented High-Level and register-Level command stream options
 - Changed High-Level command stream display to show the name of the
command
 - Fixed an issue with some operators not being displayed by the
CLI option --verbose-operators
 - Changed an unneeded print in pass packing to a more useful assertion

Change-Id: I9d53f19f4e32d0478209bc964724c27c935f66d6
Signed-off-by: Tim Hall's avatarTim Hall <tim.hall@arm.com>
      cd03504c
    • Tim Hall's avatar
      MLBEDSW-7656: Update Python versions in README · 4bd28aa1
      Tim Hall authored 
      

 - Added Python support information
 - Clarified TensorFlow support information
 - Updated Requires-Python version to 3.8

Change-Id: Iab38a2f4480e58a1bd36d5055342c4bf7379dd09
Signed-off-by: Tim Hall's avatarTim Hall <tim.hall@arm.com>
      4bd28aa1
  35. Aug 07, 2023
  37. Aug 04, 2023
  39. Jul 31, 2023
  41. Jul 24, 2023
  43. Jul 12, 2023
    • Tim Hall's avatar
      MLBEDSW-7756: MLCE: Grouped convolutions runtime problem · 9cf63a36
      Tim Hall authored 
      

 - Added graph optimiser function to convert convolution groups into
a split followed by separate convolutions and then a concat
 - Added semantic check for convolution groups
 - Added unit tests for convolution groups semantic checks
 - Fixed a minor typing issue with test_constraint_stride_range

Change-Id: I78ade408aa23469a79c9f517c4751da8619b77a9
Signed-off-by: Tim Hall's avatarTim Hall <tim.hall@arm.com>
      9cf63a36
  45. Jul 11, 2023
  47. Jul 10, 2023
    • Johan Alfvén's avatar
      MLBEDSW-7833: MLCE: Fixed output diff for reshape op · 8df91f5b
      Johan Alfvén authored and Tim Hall's avatar Tim Hall committed 
      

- In order to reduce memory usage, the live range mechanism have logic
to check if the ifm tensor can be reused for the ofm tensor for certain
operators

- In this failing test case, the input to the reshape/memcpy operator
has more than one consumer and this results in a faulty memory overwrite
since there are missing logic that should check the ifm consumers for
the memcpy operator

- The fix is to add the missing logic that ifm can only have one consumer

Change-Id: I2184c0f905b554f648c9732734098509e23b537c
Signed-off-by: Johan Alfvén's avatarJohan Alfven <johan.alfven@arm.com>
      8df91f5b
    • William Isaksson's avatar
      MLBEDSW-7752: setting query shapes to Shape4D(0) · 52afbbbe
      William Isaksson authored and Tim Hall's avatar Tim Hall committed 
      

Changes query initialization shapes to Shape4D(0,0,0,0) = [0,0,0,0]
instead of Shape4D(0) = [0,1,1,1]. The [0,1,1,1] tensors would affect
performance estimates and are not real.


Change-Id: Ic83b6f6a70c0c904b500f62756e1e125c99856c6
Signed-off-by: William Isaksson's avatarWilliam Isaksson <william.isaksson@arm.com>
      52afbbbe
  49. Jul 06, 2023
    • Tim Hall's avatar
      MLBEDSW-7832: test_tflite_model_semantic converting array to scalar · 762d3acb
      Tim Hall authored 
       - The problem is that the axis value can be either a scalar or an
array containing a single element
 - The solution is to check the length of the shape because the size
attribute returns the same value for both cases
 - This did not show up before because pytest warnings were not being
treated as errors
 - Removed pre-commit pytest option that caused tests to be searched for
from the root directory
 - Updated pyproject.toml pytest options to explicitly specify the test
directories, and to treat warnings as errors

Change-Id: I037054768e5c34f253b6062eadba1c3419ff65e4
      762d3acb
  51. Jun 28, 2023
  53. Jun 20, 2023
  55. Jun 19, 2023
  57. Jun 16, 2023