Skip to content
This project is mirrored from https://review.mlplatform.org/tosa/specification. Pull mirroring updated .
  1. Mar 23, 2021
  2. Mar 18, 2021
  3. Mar 16, 2021
    • Eric Kunze's avatar
      Add 8-bit support to TABLE · 71b02b19
      Eric Kunze authored
      
      
      It is challenging to work back from the RESCALE->TABLE->RESCALE
      for a simple 8-bit lookup. This adds support to TOSA for a
      simple direct 256 entry 8-bit lookup table for the TABLE operator.
      
      The 16-bit interpolated table is still required for a conformant
      implementation.
      
      Change-Id: I0ef218444f0b57b880aa8d1c7e96efedae72eb53
      Signed-off-by: Eric Kunze's avatarEric Kunze <eric.kunze@arm.com>
      71b02b19
    • Eric Kunze's avatar
      Note limits on axis parameter · 67ed0385
      Eric Kunze authored
      
      
      This also adds a restriction on CONCAT to have all inputs be the
      same rank.
      
      Update the ARGMAX pseudocode to handle corner cases properly and
      use axis/rank consistent with other operators.
      
      Signed-off-by: Eric Kunze's avatarEric Kunze <eric.kunze@arm.com>
      Change-Id: I941ac172ee79424ac04d562cedebb75da76659a5
      67ed0385
  4. Mar 11, 2021
    • Eric Kunze's avatar
      Make pseudocode block tags consistent · 839830a7
      Eric Kunze authored
      
      
      They are all now [source,c++] and use ---- as the
      delimiter.
      
      Also made the pseudocode indentation consistent (4 spaces)
      in the introduction sections.
      
      Signed-off-by: Eric Kunze's avatarEric Kunze <eric.kunze@arm.com>
      Change-Id: I1287a60cfeae8b7e3c63b9fc8885b1ebf3d3386d
      839830a7
    • Eric Kunze's avatar
      Adjust pseudocode types to account for zero point · 8e4a9d33
      Eric Kunze authored
      When reading tensor values with zero point, the returned value
      has one more bit than the original to account for zero point.
      
      Update cases of apply_clip to properly represent the types
      involved.
      
      Change-Id: I60c17b1b244c34b4f04f042807936ae0f282ce93
      8e4a9d33
  5. Mar 08, 2021
    • Eric Kunze's avatar
      Consistency cleanup · 1e9ba65f
      Eric Kunze authored
      
      
      Attempt to get consistent across the pseudocode.
      Change the data types to all be intN_t instead of some cases of intN.
      Use float_t as the general floating point data type.
      Be consistent on use of the term "floating-point"
      Move general pseudocode helpers to their own section.
      
      Change-Id: Ie77666cd3ee438c71f39c62b9c424fe687b0bb51
      Signed-off-by: Eric Kunze's avatarEric Kunze <eric.kunze@arm.com>
      1e9ba65f
  6. Mar 03, 2021
  7. Feb 09, 2021
  8. Feb 05, 2021
  9. Jan 27, 2021
    • Eric Kunze's avatar
      Update elementwise operator overview · f8bd5864
      Eric Kunze authored
      
      
      Elementwise operators no longer scale their inputs to a
      common range. The elementwise introductory section reflected
      the old behavior.
      
      Also clear up some language on the unary functions.
      
      Signed-off-by: Eric Kunze's avatarEric Kunze <eric.kunze@arm.com>
      Change-Id: I86bf9da8b51e9a64e4fe6766e01f0c35d43d805a
      f8bd5864
  10. Jan 26, 2021
  11. Jan 25, 2021
    • Eric Kunze's avatar
      Remove aint8 type · 2dce0d02
      Eric Kunze authored
      
      
      The aint8 type represented an asymmetrically quantized range.
      With the change to move scaling fully into the RESCALE operator,
      the aint8 and int8 types didn't have significant differences in
      their use with operators. Unifying to a single data type makes
      things simpler.
      
      Signed-off-by: Eric Kunze's avatarEric Kunze <eric.kunze@arm.com>
      Change-Id: I3ee89cbdafdc61293ef1f1bec52398d744e716a1
      2dce0d02
  12. Jan 13, 2021
  13. Dec 16, 2020
    • Eric Kunze's avatar
      Disambiguate scale_t in RESIZE command · e5d22a77
      Eric Kunze authored
      scale_t is already defined globally for the rescaling.
      
      Move pseudocode for count_leading_zeros to introduction,
      and use it in the implementation of CLZ.
      
      Change-Id: I2453bce93b6dd25e870b8a010fc62af4c001cbf9
      e5d22a77
  14. Nov 30, 2020
    • Eric Kunze's avatar
      Fix minor spec issues · 5d60c713
      Eric Kunze authored
      Profile for EXP operator
      TANH table formatting
      Remove dilation from TRANSPOSE_CONV2D
      
      Change-Id: I2a265fecfb25ad997fdc992ad8542000014ac7c0
      5d60c713
  15. Nov 13, 2020
  16. Nov 10, 2020
  17. Nov 06, 2020
  18. Nov 05, 2020
  19. Nov 03, 2020
  20. Oct 29, 2020
  21. Oct 28, 2020
  22. Oct 27, 2020
  23. Oct 22, 2020
  24. Oct 16, 2020
  25. Oct 14, 2020
  26. Oct 13, 2020
  27. Oct 12, 2020
  28. Oct 08, 2020
  29. Oct 02, 2020
  30. Oct 01, 2020
Loading