Closed geminird closed 4 years ago
tf.variance_scaling_initializer()
. Please refer to this issue SciSharp/TensorFlow.NET#356.@Oceania2018 open soruce python project at here https://github.com/jinxiwang/ocr_TDR,I'm try to use tensorflow.net implemet it. the method url at here https://github.com/jinxiwang/ocr_TDR/blob/master/test_crnn.py, line :32,another method https://github.com/jinxiwang/ocr_TDR/blob/master/crnn.py line 46,57,112。
@geminird I'm saying where is your .net version is? We'd like to help if your .net version is also open source.
@Oceania2018 The. Net version is just beginning. Because I started to learn tensorflow, as a learning sample, I can't rewrite program several problems, so I came to help. After the later program is sure to run, I will open it up on GitHub.
@Oceania2018 I pull it, see here :https://github.com/geminird/TfConsoleApp
@Oceania2018 网络不太好,没看到您的头像,您应该也是华人吧。这个是当作学习示例,目前改写了预测代码,程序跑不通。主要在CRNN类的74行,是参照BUG412修改的,但是还是不行,也把redis版本升级到1.15.0了,照样运行不了,不知道怎么解决,估计是那些匹配不到的函数替换错了。
@geminird I've raised issue on your repo: https://github.com/geminird/TfConsoleApp/issues/1
Hello, In my project, need use the ”tf.nn.ctc_greedy_decoder” and "tf.sparse_tensor_to_dense" functions , how to do in tensorflow.net api.
and can I use the "Tensorflow.Operations.rnn.static_bidirectional_rnn" funciton instead of "tf.nn.bidirectional_dynamic_rnn" function?
use "tf.variance_scaling_initializer()" instead of "tf.contrib.layers.variance_scaling_initializer()"?