Commits

Branch selector

 master

User selector

All users

Datepicker

All time

Commit History

Commits on Nov 28, 2023

• Update load_weights.py

  parssifal

  authored
  69a547c

  Copy full SHA for 69a547c

Commits on Jun 15, 2021

• Merge pull request #151 from bonlime/dependabot/pip/tensorflow-2.5.0rc0

  Show description for e32f66b
  bonlime

  authored
  e32f66b

  Copy full SHA for e32f66b

• Merge pull request #150 from vinhill/master

  Show description for 03f338f
  bonlime

  authored
  03f338f

  Copy full SHA for 03f338f

• Merge pull request #148 from YuriyKortev/replace_lambdas

  Show description for 3486fa0
  bonlime

  authored
  3486fa0

  Copy full SHA for 3486fa0

Commits on May 21, 2021

• Bump tensorflow from 2.0.0a0 to 2.5.0rc0

  Show description for f33ddd0
  dependabot[bot]

  authored
  f33ddd0

  Copy full SHA for f33ddd0

Commits on May 15, 2021

• Update model.py

  vinhill

  authored
  c5167e5

  Copy full SHA for c5167e5

Commits on Apr 1, 2021

• replaced lambdas

  YuriyKortev

  committed
  b3c9dfd

  Copy full SHA for b3c9dfd

Commits on Apr 23, 2020

• Merge pull request #141 from brianmanderson/fixing_RLock

  Show description for 1ff00aa
  bonlime

  authored
  1ff00aa

  Copy full SHA for 1ff00aa

Commits on Apr 22, 2020

• del tf2 part

  brianmanderson

  committed
  80b1de3

  Copy full SHA for 80b1de3

• need to pass the size as an int, not tensor

  brianmanderson

  committed
  1757ea0

  Copy full SHA for 1757ea0

• rollback to og code

  brianmanderson

  committed
  c67142a

  Copy full SHA for c67142a

• made return_model a def, easier to debug by moving it around

  brianmanderson

  committed
  aab2f2e

  Copy full SHA for aab2f2e

• update to tf api method of tensor.shape)

  brianmanderson

  committed
  fa42573

  Copy full SHA for fa42573

Commits on Apr 19, 2020

• going to use tf.train.experimental.enable_mixed_precition_graph_rewrite(opt)

  brianmanderson

  committed
  b05d9dc

  Copy full SHA for b05d9dc

Commits on Apr 17, 2020

• fix resize_w_pad

  brianmanderson

  committed
  5978c9b

  Copy full SHA for 5978c9b

• resize_w_pad

  brianmanderson

  committed
  3341b12

  Copy full SHA for 3341b12

• use compat image_resize, with align_corners=True

  brianmanderson

  committed
  2675dab

  Copy full SHA for 2675dab

Commits on Apr 16, 2020

• update for tf version

  brianmanderson

  committed
  c44e9f8

  Copy full SHA for c44e9f8

• update

  brianmanderson

  committed
  26c0ab8

  Copy full SHA for 26c0ab8

• keep old verison too

  brianmanderson

  committed
  38e5cf5

  Copy full SHA for 38e5cf5

• Merge pull request #1 from brianmanderson/tf2_branch

  Show description for de8ff33
  brianmanderson

  authored
  de8ff33

  Copy full SHA for de8ff33

• update to tf2

  brianmanderson

  committed
  5bd4f5c

  Copy full SHA for 5bd4f5c

• updated to latest tf2 style

  brianmanderson

  committed
  79f406e

  Copy full SHA for 79f406e

Commits on Apr 14, 2020

• back track, can't identify lambda

  brianmanderson

  committed
  2860c1d

  Copy full SHA for 2860c1d

Commits on Mar 24, 2020

• added dropout back

  brianmanderson

  committed
  cec96f0

  Copy full SHA for cec96f0

• update to elu, remove dropout

  brianmanderson

  committed
  6407677

  Copy full SHA for 6407677

• update to be w elu and no drop

  brianmanderson

  committed
  df9a133

  Copy full SHA for df9a133

Commits on Mar 21, 2020

• reverted back to the paper arch, just updated to allow to run on keras, will branch and change activations to elu on my own

  brianmanderson

  committed
  1e7c221

  Copy full SHA for 1e7c221

• turn all relu -> elu, remove dropout, found it improved stability while training

  brianmanderson

  committed
  ac74418

  Copy full SHA for ac74418

• remove loading tensorflow

  brianmanderson

  committed
  6836002

  Copy full SHA for 6836002

• finish updating to be original implementation, with keras usage, and removing problem of model save w RLock

  brianmanderson

  committed
  1a10c8f

  Copy full SHA for 1a10c8f

• update to use Upsampling layer instead of tf.resize, trying to get around the RLock issue

  brianmanderson

  committed
  7175af3

  Copy full SHA for 7175af3

Commits on Mar 20, 2020

• update to be elu instead of relu, drop dropout

  brianmanderson

  committed
  a13b3e5

  Copy full SHA for a13b3e5

• still a mystery.. RLock only seems to happen when the image.resize is used, but the keras one gives poorer performance

  brianmanderson

  committed
  ccd11a6

  Copy full SHA for ccd11a6

• update to be fully keras, not needing to load relu6

  brianmanderson

  committed
  56b8021

  Copy full SHA for 56b8021

Pagination

Previous

Next