qixiang109
qixiang109
@sonack I agree with you. We need padding filter_size-1 zeros to the left with each layer.
@sonack 不好意思很久没看这边，我还没具体做过gated cnn语言模型的实验，但是我很确定补0的方式
is it possible to install 10.13.6 on Z590？i need to run cuda so the latest compatible system version is 10.13.6…
guess as the log-normal's natural, smaller sigma on the Gaussian prior would give your smoother topic proportions. However this implement seems do not allow for a configurable sigma——it is hard-coded...
> Found a way to update the trie tree node We could find the way to initailize the trie tree is adding a list of words to the `marisa_trie.Trie` Just...
The same question. Does a 4-layer 4-spatial filter 2-strides convolution naturally leads to a 70*70 patch discriminator?