-
Occurs in Graph execution (not Eager) with a `return_sequences=True` followed by `return_sequences=False` (but not either on its own) custom RNN layer; the layer involves `nn.moments` and `nn.batch_no…
-
Using pytorch 1.5, cuda 10.1, python 3.7.7 in a clean virtual environment, using the haste-0.4 code under releases.
It properly compiles libhaste.a, then through tracing the makefile it attemps to …
-
P155，介绍了堆叠循环神经网络。但是RNN和LSTM在深层堆叠方面效果较差，一般只能堆叠三层，超过三层反而会带来效果下降。独立循环神经网络（Independently Recurrent Neural Network， IndRNN）可以堆叠深层网络，并且达到非常好的效果。其堆叠形式可以完全参考CNN现有的普通堆叠，残差堆叠，密集堆叠等多种方式。
P149（长程依赖问题），IndRNN完美…
-
Looking forward to reply~~
-
In particular IndRNNs and likely RNNs, maybe also LSTMs - looking into it, will fix
-
hello, could you help share some ideas about how to reimplement the idea on the action recognition dataset?
I tried to implement this, but the performance is not good, hope you can give some instruc…
-
Sorry to open an issue here. However, I think that you are an expert on this topic, and IndRNNCell related upgrade in TensorFlow 1.10.0 might also be created by you.
I am trying to apply relu activ…
-
Hi,
Thanks for the wonderful work. I tried to reproduce the results for Action Recognition task on NTU RGB+D dataset on the subject split, by running the provided command.
- python -u Indrnn_actio…
-
hello, thanks for your excellent work. I notice that you implementation on action recognition is different from the paper formulated. I want to know why you have those changes in your codebase?
ps: y…
-
I'm using Keras 2.1.5 but i am getting this error when I try to run IndRNN:
ImportError Traceback (most recent call last)
in ()
11 from keras.legacy import int…