8000 Commits · RonB/models · GitHub

More Web Proxy on the site http://driver.im/

Commits

Commits on Mar 25, 2020

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

reedwm
authored and
tensorflower-gardener
committed

Commits on Feb 25, 2020

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

reedwm
authored and
tensorflower-gardener
committed

Commits on Nov 1, 2019

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

reedwm
authored and
tensorflower-gardener
committed

Commits on Oct 16, 2019

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

reedwm
authored and
tensorflower-gardener
committed

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

reedwm
authored and
tensorflower-gardener
committed

Commits on Oct 9, 2019

Simply LayerNorm mixed precision logic.

reedwm
authored and
tensorflower-gardener
committed

Commits on Sep 9, 2019

Unexpose some flags from models which do not use them.

reedwm
authored and
tensorflower-gardener
committed

Commits on Sep 4, 2019

Unexpose some flags from models which do not use them.

reedwm
authored and
tensorflower-gardener
committed

Commits on Aug 30, 2019

Fix all lint errors in official/vision/image_classification/

reedwm
authored and
tensorflower-gardener
committed
Use tf.compat.v2 when accessing policies.

reedwm
authored and
tensorflower-gardener
committed
Fix bug where dynamic loss scaling was broken

reedwm
authored and
tensorflower-gardener
committed

Commits on Aug 29, 2019

Use new mixed_float16 policy for resnet.

reedwm
authored and
tensorflower-gardener
committed
Enable V2 behavior in transformer_main.py.

reedwm
authored and
tensorflower-gardener
committed

Commits on Aug 26, 2019

Unexpose some flags from models which do not use them.

reedwm
authored and
tensorflower-gardener
committed

Commits on Aug 23, 2019

Unexpose some flags from models which do not use them.

reedwm
authored and
tensorflower-gardener
committed

Commits on Aug 21, 2019

Use tf.float32 instead of "float32"
reedwm
committed
Internal change

reedwm
authored and
tensorflower-gardener
committed

Commits on Aug 20, 2019

Use new mixed_float16 policy for transformer.
reedwm
committed

Commits on Aug 19, 2019

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

reedwm
authored and
tensorflower-gardener
committed
Fix transformer_main.py.

reedwm
authored and
tensorflower-gardener
committed

Commits on Aug 16, 2019

Show flags when --help is specified in resnet.
reedwm
committed
Fix transformer fp16 test

reedwm
authored and
tensorflower-gardener
committed

Commits on Aug 8, 2019

Fix fp16 Transformer model. (#7402)
reedwm
authored

Commits on Jun 21, 2019

Fix Transformer Perfzero issue with fp16 (#7074)

reedwm
authored and
tfboyd
committed

Commits on Jun 19, 2019

Add mixed precision support to Transformer (#7011)
reedwm
authored

Commits on Jun 6, 2019

Commits on May 18, 2019

Include flags when reporting benchmark. (#6809)

reedwm
authored and
tfboyd
committed

Commits on May 3, 2019

Add graph rewrite convergence benchmark (#6712)
reedwm
authored

Commits on May 2, 2019

Add graph rewrite benchmarks (#6708)
reedwm
authored

Commits on May 1, 2019

Add --fp16_implementation option. (#6703)
reedwm
authored

Commits on Apr 5, 2019

Fix dynamic loss scaling crash in benchmarks (#6532)
reedwm
authored

Commits on Apr 4, 2019

Add benchmarks for dynamic loss scaling. (#6520)

reedwm
authored and
tfboyd
committed

Commits on Apr 3, 2019

Add dynamic loss scaling support (#6518)
reedwm
authored

Commits on Mar 7, 2019

Add fp16 to keras benchmarks (#6314)

reedwm
authored and
tfboyd
committed

0