suiyuan2009

https://github.com/crownpku/Information-Extraction-Chinese/blob/976a528cb22565ce3b72a4a2b2837a412af4881b/RE_BGRU_2ATT/initial.py#L191

global_step should be protected by a lock?

4

I have met stuck problem when running `tf_cnn_benchmarks.py` in distributed mode, I think `global_step` should be protected by a lock in [this line](https://github.com/tensorflow/benchmarks/blob/master/scripts/tf_cnn_benchmarks/tf_cnn_benchmarks.py#L934).

add support for mock data input to imagenet example

1

in order to test the performance conveniently, especially distributed training, it would be great to add an option to allow mock data input, which means directly random some data in...

in my test NHWC format is 25% slow than NCHW on resnet101, and non fused batch normalization is very slow and uses much more memory than fused bn.

get errors when build ICTCLASTest.csproj

1

ubuntu 15.04, use xbuild(mono) when I run `xbuild ICTCLASTest.csproj` in `NLPIR/NLPIR SDK/NLPIR-ICTCLAS/projects/ICTCLAS_C#`, I get ``` XBuild Engine Version 12.0 Mono, Version 3.2.8.0 Copyright (C) 2005-2013 Various Mono authors Build started...