Commits

Branch selector

 master

User selector

wanxinwx

Datepicker

All time

Commit History

Commits on Oct 6, 2021

• Changed comparison statement so that int and IntEnum field are properly compared.

  Show description for 2ac17ab
  

  wanxinwx

  authored and

  copybara-github

  committed
  2ac17ab

Commits on May 4, 2021

• Changed initialization for codebook and encoding matrices for k-means compression op for faster graph construction.

  Show description for bf2684c
  

  wanxinwx

  authored and

  copybara-github

  committed
  bf2684c

Commits on Apr 19, 2021

• Added spec argument for SimhashApplyCompression.apply_compression() method.

  Show description for 3b89d40
  

  wanxinwx

  authored and

  copybara-github

  committed
  3b89d40

Commits on Apr 5, 2021

• added an option to disable compression summaries.

  Show description for a77eab3
  

  wanxinwx

  authored and

  copybara-github

  committed
  a77eab3

Commits on Mar 30, 2021

• Added compression/pruning support for embedding layers in the SimpleEmbeddingLayer class.

  Show description for 674675c
  

  wanxinwx

  authored and

  copybara-github

  committed
  674675c

Commits on Mar 23, 2021

• Set update_ops list to be empty for compression_option 9 (InputOutputCompression).

  Show description for b76546e
  

  wanxinwx

  authored and

  copybara-github

  committed
  b76546e

Commits on Mar 8, 2021

• Added dependency for graph compression library.

  Show description for 9bd2492
  

  wanxinwx

  authored and

  copybara-github

  committed
  9bd2492

Commits on Feb 24, 2021

• Added weight_name argument in the PruningOp.ApplyPruning() method.

  Show description for 470cf5a
  

  wanxinwx

  authored and

  copybara-github

  committed
  470cf5a

Commits on Feb 20, 2021

• Added methods to support unified pruning/compression interface.

  Show description for 102b6a5
  

  wanxinwx

  authored and

  copybara-github

  committed
  102b6a5

Commits on Feb 18, 2021

• Removed dependency on tensorflow.contrib for HParams.

  Show description for c374ee4
  

  wanxinwx

  authored and

  copybara-github

  committed
  c374ee4

Commits on Feb 9, 2021

• Added group pruning option.

  Show description for e245e90
  

  wanxinwx

  authored and

  copybara-github

  committed
  e245e90

Commits on Nov 5, 2020

• Added ModelPruningListener and ModelPruningHook to allow running Python pruning updates in tf.estimator.Estimator.

  Show description for 007d1ce
  

  wanxinwx

  authored and

  copybara-github

  committed
  007d1ce

Commits on Oct 6, 2020

• Added do_transpose option for KmeansPruningCompressionOp and reformatted the get_mix_operator function.

  Show description for edacc4b
  

  wanxinwx

  authored and

  copybara-github

  committed
  edacc4b

Commits on Sep 23, 2020

• Added update_option 2 and make it the default for KmeansPruningCompressionOp.

  Show description for 486edb1
  

  wanxinwx

  authored and

  copybara-github

  committed
  486edb1

Commits on Aug 6, 2020

• Added KmeansPruningCompressionOp for pruning + kmeans vector quantization.

  Show description for 52620de
  

  wanxinwx

  authored and

  copybara-github

  committed
  52620de

Commits on Mar 26, 2020

• Added methods to support compression for customized layers that handles variable creation.

  Show description for 48424ff
  

  wanxinwx

  authored and

  copybara-github

  committed
  48424ff

Commits on Mar 16, 2020

• Fixed global_step issue and corrected the pruning interface example.

  Show description for 1be2b36
  

  wanxinwx

  authored and

  copybara-github

  committed
  1be2b36

Commits on Feb 21, 2020

• Added options to retrieve update_ops from collections for compression.

  Show description for 406b3b6
  

  wanxinwx

  authored and

  copybara-github

  committed
  406b3b6

Commits on Feb 19, 2020

• Changed hparams.get_operator_hparams to directly accessing hparams fields.

  Show description for 11c9bf3
  

  wanxinwx

  authored and

  copybara-github

  committed
  11c9bf3

Commits on Feb 18, 2020

• Fixed typo in end_compression_step argument.

  Show description for 303fa7f
  

  wanxinwx

  authored and

  copybara-github

  committed
  303fa7f

Commits on Feb 6, 2020

• added supports for compression_lib.

  Show description for 94d1859
  

  wanxinwx

  authored and

  copybara-github

  committed
  94d1859

Commits on Feb 3, 2020

• Added kmeans_quantize.py and kmeans_quantize_test.py, as utilities of kmeans compressor. Added dictionary learning compression op.

  Show description for 369145f
  

  wanxinwx

  authored and

  copybara-github

  committed
  369145f

Commits on Feb 1, 2020

• added supports for compression_lib.

  Show description for 0ca8615
  

  wanxinwx

  authored and

  copybara-github

  committed
  0ca8615

Commits on Jan 26, 2020

• Migrated to TensorFlow 2.0.

  Show description for 6354069
  

  wanxinwx

  authored and

  copybara-github

  committed
  6354069

Commits on Jan 21, 2020

• Added optimization for the dictionary learning function to handle large input data.

  Show description for e25effb
  

  wanxinwx

  authored and

  copybara-github

  committed
  e25effb

Commits on Dec 16, 2019

• Changed import names for the cifar10 example.

  Show description for f43cb7f
  

  wanxinwx

  authored and

  copybara-github

  committed
  f43cb7f

Commits on Dec 12, 2019

• Changed import names.

  Show description for 51345f8
  

  wanxinwx

  authored and

  copybara-github

  committed
  51345f8

Commits on Nov 25, 2019

• Added block pruning for gradient based pruning options.

  Show description for d2cb5b4
  

  wanxinwx

  authored and

  copybara-github

  committed
  d2cb5b4

Commits on Oct 30, 2019

• Added pruning_interface for easy access to the pruning APIs.

  Show description for 7b5f773
  

  wanxinwx

  authored and

  copybara-github

  committed
  7b5f773