Code

Leaky Quantized ReLU fix#961

Code

Closed

DrWatt wants to merge 34 commits intofastmachinelearning:mainfastmachinelearning/hls4ml:mainfrom DrWatt:qrelu_fixDrWatt/hls4ml:qrelu_fixCopy head branch name to clipboard

+82-8

Commits

Commits on Mar 29, 2023

• First attempt in making a layer generator for hls4ml to make adding a new layer easier

  DrWatt

  committed
  4e483fc

  Copy full SHA for 4e483fc

Commits on Apr 3, 2023

• Added longnamelayer to associate hls layer to TF layer and added type and default values to general attributes

  DrWatt

  committed
  0e973e6

  Copy full SHA for 0e973e6

• Merge branch 'fastmachinelearning:main' into master

  DrWatt

  authored
  ebbac33

  Copy full SHA for ebbac33

Commits on Jan 18, 2024

• Merge branch 'fastmachinelearning:main' into master

  DrWatt

  authored
  e1e9157

  Copy full SHA for e1e9157

• Fixed a missing casting in vivado_accelerator and added a skip for linear activation embedded in qkeras layers without activation in k_to_hls converter

  DrWatt

  committed
  49a30fe

  Copy full SHA for 49a30fe

• Fixed missing _ap_ in typestring sometimes not supported by the compiler

  DrWatt

  committed
  e73ddc9

  Copy full SHA for e73ddc9

Commits on Jan 24, 2024

• Fixed issue with quantized_relu activation whose implementation did not produce leakyrelu activations when negative slope is provided

  DrWatt

  committed
  6db982a

  Copy full SHA for 6db982a

• Removed a couple of prints used in debugging

  DrWatt

  committed
  6ea8ab0

  Copy full SHA for 6ea8ab0

• Merge branch 'fastmachinelearning:main' into master

  DrWatt

  authored
  a13c978

  Copy full SHA for a13c978

Commits on Jan 25, 2024

• testing test

  DrWatt

  committed
  6640169

  Copy full SHA for 6640169

• pre-commit ran manually on all files

  DrWatt

  committed
  6028bb6

  Copy full SHA for 6028bb6

• Fixed issues arised due to an old version of hls4ml used when writing the layer templates

  DrWatt

  committed
  6c72dd9

  Copy full SHA for 6c72dd9

• Fixed hls header name

  DrWatt

  committed
  61ae795

  Copy full SHA for 61ae795

Commits on Jan 26, 2024

• first hls compilation errors solved in new layer template

  DrWatt

  committed
  7b4047d

  Copy full SHA for 7b4047d

• Readme added for the New Layer tool

  DrWatt

  committed
  235aa5a

  Copy full SHA for 235aa5a

• Added notebook for testing quickly a new layer

  DrWatt

  committed
  99afa33

  Copy full SHA for 99afa33

• Changed name of directory

  DrWatt

  committed
  7dd760d

  Copy full SHA for 7dd760d

• Old directory deleted

  DrWatt

  committed
  11371c5

  Copy full SHA for 11371c5

• Missing outputshape description in Readme

  DrWatt

  committed
  eed800d

  Copy full SHA for eed800d

Commits on Jan 29, 2024

• Removing stuff that do not concern the bug fix

  DrWatt

  committed
  fde9841

  Copy full SHA for fde9841

• Uncommented line in profiling

  DrWatt

  committed
  372d35d

  Copy full SHA for 372d35d

• Test suite added for quantized leaky relu

  DrWatt

  committed
  928f383

  Copy full SHA for 928f383

• Changed output_dir of tests to follow the same configuration as the others test scripts

  DrWatt

  committed
  3b561b4

  Copy full SHA for 3b561b4

Commits on Jan 30, 2024

• Merge branch 'fastmachinelearning:main' into qrelu_fix

  DrWatt

  authored
  da65388

  Copy full SHA for da65388

Commits on Feb 8, 2024

• Removed 'ap_' string from FixedPrecisionType and fixed compiling error regarding ap_fixed types temporarily directly in the vivado accelerator writer

  DrWatt

  committed
  f1d187d

  Copy full SHA for f1d187d

• Added if clause in the quantized_relu config which makes it signed or not according to the negative slope

  DrWatt

  committed
  700163b

  Copy full SHA for 700163b

Commits on Feb 9, 2024

• Renamed QLeakyReLU to LeakyReLU and subsequent changes

  DrWatt

  committed
  3f48509

  Copy full SHA for 3f48509

• Reversed changes regarding vivado_accelerator backend and writer

  DrWatt

  committed
  d209f5b

  Copy full SHA for d209f5b

• Merge branch 'fastmachinelearning:main' into qrelu_fix

  DrWatt

  authored
  46a6e9b

  Copy full SHA for 46a6e9b

• changed description string of qrelu test

  DrWatt

  committed
  26f32a1

  Copy full SHA for 26f32a1

• Name of qleaky relu test changed

  DrWatt

  committed
  d4642e7

  Copy full SHA for d4642e7

• Name of qleaky relu test changed

  DrWatt

  committed
  b1a4472

  Copy full SHA for b1a4472

Commits on Feb 19, 2024

• Merge branch 'fastmachinelearning:main' into qrelu_fix

  DrWatt

  authored
  ecec240

  Copy full SHA for ecec240

Commits on Mar 27, 2024

• Merge branch 'main' into qrelu_fix

  DrWatt

  authored
  34c8070

  Copy full SHA for 34c8070