Commits

Branch selector

 master

User selector

reedwm

Datepicker

All time

Commit History

Commits on Apr 9, 2021

• Remove dynamic_loss_scale argument to define_performance.

  Show description for e353e4e
  

  reedwm

  authored and

  tensorflower-gardener

  committed
  e353e4e

  Copy full SHA for e353e4e

Commits on Apr 6, 2021

• Change default value of use_experimental_api to False.

  Show description for 417aa07
  

  reedwm

  authored and

  tensorflower-gardener

  committed
  417aa07

  Copy full SHA for 417aa07

• Pass use_experimental_api=True to mixed precision functions.

  Show description for 9fe8522
  

  reedwm

  authored and

  tensorflower-gardener

  committed
  9fe8522

  Copy full SHA for 9fe8522

• Use tf.compat.v1 version of enable_mixed_precision_graph_rewrite.

  Show description for 5cdbcac
  

  reedwm

  authored and

  tensorflower-gardener

  committed
  5cdbcac

  Copy full SHA for 5cdbcac

Commits on Apr 5, 2021

• Use nonexperimental LSO API in base_task.py.

  Show description for cc12499
  

  reedwm

  authored and

  tensorflower-gardener

  committed
  cc12499

  Copy full SHA for cc12499

Commits on Mar 3, 2021

• Pass use_experimental_api=True to mixed precision functions.

  Show description for 4ab6f1b
  

  reedwm

  authored and

  tensorflower-gardener

  committed
  4ab6f1b

  Copy full SHA for 4ab6f1b

Commits on Feb 25, 2021

• Fix issue where distribution_strategy="off" did not work.

  Show description for 4864e79
  

  reedwm

  authored and

  tensorflower-gardener

  committed
  4864e79

  Copy full SHA for 4864e79

Commits on Nov 20, 2020

• Internal change

  Show description for 1b5a4c9
  

  reedwm

  authored and

  tensorflower-gardener

  committed
  1b5a4c9

  Copy full SHA for 1b5a4c9

Commits on Nov 17, 2020

• Internal change

  Show description for 68068c0
  

  reedwm

  authored and

  tensorflower-gardener

  committed
  68068c0

  Copy full SHA for 68068c0

Commits on Nov 14, 2020

• Internal change

  Show description for f409e4d
  

  reedwm

  authored and

  tensorflower-gardener

  committed
  f409e4d

  Copy full SHA for f409e4d

Commits on Nov 13, 2020

• Internal change

  Show description for e1d7daf
  

  reedwm

  authored and

  tensorflower-gardener

  committed
  e1d7daf

  Copy full SHA for e1d7daf

Commits on Oct 17, 2020

• Do not access self.embeddings.dtype.

  Show description for 126ce65
  

  reedwm

  authored and

  tensorflower-gardener

  committed
  126ce65

  Copy full SHA for 126ce65

Commits on May 28, 2020

• Use float32 activation in Transformer.

  Show description for 94b1efc
  

  reedwm

  authored and

  tensorflower-gardener

  committed
  94b1efc

  Copy full SHA for 94b1efc

Commits on Mar 25, 2020

• Check if there are GPUs instead of if TF is built with CUDA support.

  Show description for ad09cf4
  

  reedwm

  authored and

  tensorflower-gardener

  committed
  ad09cf4

  Copy full SHA for ad09cf4

Commits on Feb 25, 2020

• Convert numpy arrays to floats using float constructor, not .item().

  Show description for ed2e3bc
  

  reedwm

  authored and

  tensorflower-gardener

  committed
  ed2e3bc

  Copy full SHA for ed2e3bc

Commits on Nov 1, 2019

• Support fp16 using tf.keras.mixed_precision in CTL resnet.

  Show description for ffa522e
  

  reedwm

  authored and

  tensorflower-gardener

  committed
  ffa522e

  Copy full SHA for ffa522e

Commits on Oct 16, 2019

• Add support for the tf.keras.mixed_precision API in NCF

  Show description for cb91369
  

  reedwm

  authored and

  tensorflower-gardener

  committed
  cb91369

  Copy full SHA for cb91369

Commits on Oct 11, 2019

• Do not show the percent symbol for accuracy, as the value is between 0 and 1, not 0 and 100

  Show description for 9f8e064
  

  reedwm

  authored and

  tensorflower-gardener

  committed
  9f8e064

  Copy full SHA for 9f8e064

Commits on Oct 9, 2019

• Simply LayerNorm mixed precision logic.

  Show description for 0257b27
  

  reedwm

  authored and

  tensorflower-gardener

  committed
  0257b27

  Copy full SHA for 0257b27

Commits on Sep 9, 2019

• Unexpose some flags from models which do not use them.

  Show description for e91c41c
  

  reedwm

  authored and

  tensorflower-gardener

  committed
  e91c41c

  Copy full SHA for e91c41c

Commits on Sep 4, 2019

• Unexpose some flags from models which do not use them.

  Show description for a85c40e
  

  reedwm

  authored and

  tensorflower-gardener

  committed
  a85c40e

  Copy full SHA for a85c40e

Commits on Aug 30, 2019

• Fix all lint errors in official/vision/image_classification/

  Show description for b9ef963
  

  reedwm

  authored and

  tensorflower-gardener

  committed
  b9ef963

  Copy full SHA for b9ef963

• Use tf.compat.v2 when accessing policies.

  Show description for a6f9945
  

  reedwm

  authored and

  tensorflower-gardener

  committed
  a6f9945

  Copy full SHA for a6f9945

• Fix bug where dynamic loss scaling was broken

  Show description for 765da42
  

  reedwm

  authored and

  tensorflower-gardener

  committed
  765da42

  Copy full SHA for 765da42

Commits on Aug 29, 2019

• Use new mixed_float16 policy for resnet.

  Show description for dcd0e7a
  

  reedwm

  authored and

  tensorflower-gardener

  committed
  dcd0e7a

  Copy full SHA for dcd0e7a

• Enable V2 behavior in transformer_main.py.

  Show description for 68c3c65
  

  reedwm

  authored and

  tensorflower-gardener

  committed
  68c3c65

  Copy full SHA for 68c3c65

Commits on Aug 26, 2019

• Unexpose some flags from models which do not use them.

  Show description for 560b3af
  

  reedwm

  authored and

  tensorflower-gardener

  committed
  560b3af

  Copy full SHA for 560b3af

Commits on Aug 23, 2019

• Unexpose some flags from models which do not use them.

  Show description for 882e51a
  

  reedwm

  authored and

  tensorflower-gardener

  committed
  882e51a

  Copy full SHA for 882e51a

Commits on Aug 21, 2019

• Use tf.float32 instead of "float32"

  reedwm

  committed
  dfcca06

  Copy full SHA for dfcca06

• Internal change

  Show description for 8856b91
  

  reedwm

  authored and

  tensorflower-gardener

  committed
  8856b91

  Copy full SHA for 8856b91

Commits on Aug 20, 2019

• Use new mixed_float16 policy for transformer.

  Show description for acdf24c
  reedwm

  committed
  acdf24c

  Copy full SHA for acdf24c

Commits on Aug 19, 2019

• Do not expose --max_train_steps in models that do not use it.

  Show description for 824ff2d
  

  reedwm

  authored and

  tensorflower-gardener

  committed
  824ff2d

  Copy full SHA for 824ff2d

• Fix transformer_main.py.

  Show description for b1188d0
  

  reedwm

  authored and

  tensorflower-gardener

  committed
  b1188d0

  Copy full SHA for b1188d0

Commits on Aug 16, 2019

• Show flags when --help is specified in resnet.

  reedwm

  committed
  a2a8ff8

  Copy full SHA for a2a8ff8

• Fix transformer fp16 test

  Show description for 7b57b6d
  

  reedwm

  authored and

  tensorflower-gardener

  committed
  7b57b6d

  Copy full SHA for 7b57b6d

Pagination

Previous

Next