Commits

Branch selector

 master

User selector

lishanok

Datepicker

All time

Commit History

Commits on Mar 23, 2022

• Internal change

  Show description for 7be2baf
  

  lishanok

  authored and

  copybara-github

  committed
  7be2baf

  Copy full SHA for 7be2baf

• Internal change

  Show description for 80b5244
  

  lishanok

  authored and

  copybara-github

  committed
  80b5244

  Copy full SHA for 80b5244

• Internal change

  Show description for de94b88
  

  lishanok

  authored and

  copybara-github

  committed
  de94b88

  Copy full SHA for de94b88

Commits on Feb 17, 2022

• Internal change

  Show description for af0ddb8
  

  lishanok

  authored and

  copybara-github

  committed
  af0ddb8

  Copy full SHA for af0ddb8

Commits on Feb 7, 2022

• Internal change

  Show description for 39dc01d
  

  lishanok

  authored and

  copybara-github

  committed
  39dc01d

  Copy full SHA for 39dc01d

Commits on Feb 1, 2022

• Internal change

  Show description for a7cebdb
  

  lishanok

  authored and

  copybara-github

  committed
  a7cebdb

  Copy full SHA for a7cebdb

Commits on Jan 25, 2022

• Internal change

  Show description for 6324363
  

  lishanok

  authored and

  copybara-github

  committed
  6324363

  Copy full SHA for 6324363

• Internal change

  Show description for 3a06de5
  

  lishanok

  authored and

  copybara-github

  committed
  3a06de5

  Copy full SHA for 3a06de5

Commits on Jan 8, 2022

• Adjust multiplier and accumulator bit width when scaling factor is auto_po2

  Show description for 4bfe8af
  

  lishanok

  authored and

  copybara-github

  committed
  4bfe8af

  Copy full SHA for 4bfe8af

Commits on May 26, 2021

• Fix qtools to assure reference energy is the same as trial in floating point model.

  Show description for 5c33b07
  

  lishanok

  authored and

  copybara-github

  committed
  5c33b07

  Copy full SHA for 5c33b07

Commits on May 7, 2021

• Fixes bug where tensor type is used in an if statement.

  Show description for a447723
  

  lishanok

  authored and

  copybara-github

  committed
  a447723

  Copy full SHA for a447723

• Fix complications where QAdaptiveActivation.integer is a tensor type.

  Show description for a67865d
  

  lishanok

  authored and

  copybara-github

  committed
  a67865d

  Copy full SHA for a67865d

Commits on Apr 21, 2021

• Add support for TFOpLambda layer in model_quantize function.

  Show description for 672c134
  

  lishanok

  authored and

  copybara-github

  committed
  672c134

  Copy full SHA for 672c134

• Internal change

  Show description for c731359
  

  lishanok

  authored and

  copybara-github

  committed
  c731359

  Copy full SHA for c731359

Commits on Apr 15, 2021

• Add support for Batchnorm folded layers in print_qstats and sparsity calculation.

  Show description for eb6e0dc
  

  lishanok

  authored and

  copybara-github

  committed
  eb6e0dc

  Copy full SHA for eb6e0dc

• Implementation of h-swish activation quantizer

  Show description for 4faff4a
  

  lishanok

  authored and

  copybara-github

  committed
  4faff4a

  Copy full SHA for 4faff4a

• Add support for float type in estimate.py

  Show description for 9ca7ec2
  

  lishanok

  authored and

  copybara-github

  committed
  9ca7ec2

  Copy full SHA for 9ca7ec2

Commits on Feb 20, 2021

• Fix misc bugs in Batchnorm folded layers and the related utility functions.

  Show description for 25e7803
  

  lishanok

  authored and

  copybara-github

  committed
  25e7803

  Copy full SHA for 25e7803

Commits on Feb 19, 2021

• Internal change

  Show description for 67e7c6b
  

  lishanok

  authored and

  copybara-github

  committed
  67e7c6b

  Copy full SHA for 67e7c6b

• Internal change

  Show description for b9c4bea
  

  lishanok

  authored and

  copybara-github

  committed
  b9c4bea

  Copy full SHA for b9c4bea

Commits on Feb 4, 2021

• Add test for converting QDepthwiseConv2DBatchnorm to QDepthwiseCon2D layers.

  Show description for ebff2bc
  

  lishanok

  authored and

  copybara-github

  committed
  ebff2bc

  Copy full SHA for ebff2bc

Commits on Jan 26, 2021

• Fix the bug where the multiplier to depthwiseconv2d kernel needs to be reshaped in order to be broadcasted properly when multiplied with the kernel. Fix model_quantize function so that bn_freeze_de…

  Show description for 530fd6c
  

  lishanok

  authored and

  copybara-github

  committed
  530fd6c

  Copy full SHA for 530fd6c

Commits on Dec 16, 2020

• Internal change

  Show description for 212d331
  

  lishanok

  authored and

  copybara-github

  committed
  212d331

  Copy full SHA for 212d331

• Implement function populate_bias_quantizer_for_folded_layers. When user set bias_quantizer=None for batchnorm folded layers, we generate the accumulator type of the layer MAC op and set it as the b…

  Show description for 2e79554
  

  lishanok

  authored and

  copybara-github

  committed
  2e79554

  Copy full SHA for 2e79554

• Convert a qkeras sequential and non-sequential model with folded layer to unfolded layer and set their weights accordingly.

  Show description for bc61890
  

  lishanok

  authored and

  copybara-github

  committed
  bc61890

  Copy full SHA for bc61890

Commits on Nov 25, 2020

• Internal change

  Show description for 671ff87
  

  lishanok

  authored and

  copybara-github

  committed
  671ff87

  Copy full SHA for 671ff87

Commits on Nov 18, 2020

• Internal change

  Show description for adcdc44
  

  lishanok

  authored and

  copybara-github

  committed
  adcdc44

  Copy full SHA for adcdc44

Commits on Oct 12, 2020

• Internal change

  Show description for 98d8310
  

  lishanok

  authored and

  copybara-github

  committed
  98d8310

  Copy full SHA for 98d8310

Commits on Oct 9, 2020

• Extend autoqkeras to support multi-head models.

  Show description for 4e5035d
  

  lishanok

  authored and

  copybara-github

  committed
  4e5035d

  Copy full SHA for 4e5035d

Commits on Oct 1, 2020

• Internal change

  Show description for e3d6778
  

  lishanok

  authored and

  copybara-github

  committed
  e3d6778

  Copy full SHA for e3d6778

Commits on Sep 25, 2020

• Internal change

  Show description for ba6c5b9
  

  lishanok

  authored and

  copybara-github

  committed
  ba6c5b9

  Copy full SHA for ba6c5b9

Commits on Aug 3, 2020

• This CL enables autoqkeras to save trained model after each block tuning. Once Autoqkeras crashes, this CL enables reading the oracle of the block to be resumed and autoqkeras tuning can be resumed…

  Show description for 12e5832
  

  lishanok

  authored and

  copybara-github

  committed
  12e5832

  Copy full SHA for 12e5832

• Depending on different tensorflow version (e.g., 2.4), layer_output.shape might be tuple instead of tensorshape. We need to use shape.as_list() or list(shape) accordingly to convert shape to list.

  Show description for 16e7c26
  

  lishanok

  authored and

  copybara-github

  committed
  16e7c26

  Copy full SHA for 16e7c26

• Internal change

  Show description for 68216cb
  

  lishanok

  authored and

  copybara-github

  committed
  68216cb

  Copy full SHA for 68216cb

Commits on Jul 17, 2020

• In pytype_strict_libarary/binary rules, direct dependencies are enforced.

  Show description for e261e97
  

  lishanok

  authored and

  copybara-github

  committed
  e261e97

  Copy full SHA for e261e97

Pagination

Previous

Next