dzhu622911

I have used POST already, so it seems POST is the only choice

I also read the Thesis and this code carefully, and found the same question with above, the group_index seems not take effect, the expected effect maybe group_index * group_size? which...

here is modified code with group_index * group_size `import numpy as np import tensorflow as tf class ClockworkRNN(object): ''' A Clockwork RNN - Koutnik et al. 2014 [arXiv, https://arxiv.org/abs/1402.3511] The...

@tomrunia , my friend zhlicen already upload the code, please review. and i found another question, will submit another issue.

sorry, should be this: ``` mask = np.zeros((self.config.num_hidden, self.config.num_hidden)) for i in range(len(self.clockwork_periods)): mask[i*self.group_size:(i+1)*self.group_size, i*self.group_size:] = 1 self.clockwork_mask = tf.constant(mask, dtype=tf.float32, name="mask") ```