Hao

icon indicating a website link https://zsdonghao.github.io

icon company Peking University icon location Beijing icon location Assistant Professor @ Peking University

Results 146 comments of Hao

For our own dataset

For customised dataset, we need to prepare the data to fit the same format here : https://github.com/zsdonghao/text-to-image/blob/master/data_loader.py#L166 BTW, this is the steps to create the vocabulary: https://github.com/wagamamaz/tensorlayer-tricks/blob/master/README.md#9-sentences-tokenization

MemoryError [solved]

it is MemoryError... you may need to solve it yourself

MemoryError [solved]

cool~ thank you I will mark it done~

How to combine the text caption and images in the generator

This project is the reimplementation of GAN-CLS, please read that paper

ch 6.4 p166 按照书中的例子输入的代码，初始cost非常小

你是自己手动把代码输入到电脑中的吗？建议直接复制每章对应的网上例子，手动输入很容易搞错的。

TensorLayer书 第一印 第九章 9.3 vocal.word_to_id('one') 输出应该为 2

209页： ```python >>> vocab.id_to_word(2) ... 'one' >>> vocab.word_to_id('one') ... 2 # 这里 ```

ch3.5 p69 line 30 ReconLayer.pretrain denoise_name not including variable_scope

书里没有```with tf.variable_scope("ae", reuse=reuse):```哦，你自己改的吗？

ch3.5 p69 line 30 ReconLayer.pretrain denoise_name not including variable_scope

能否把代码都贴上来，我看了下 ReconLayer的源码，没发现问题～

ch3.5 p69 line 30 ReconLayer.pretrain denoise_name not including variable_scope

I test the code after adding `variable_scope `, it works well.

ch 8.3 p198 line 2 all_images没有定义

hi, 以这个为准：https://github.com/tensorlayer/dcgan/blob/master/main.py