entalent / memcap Goto Github PK

for example.

clf_textcnn_word2idx_path = '/media/wentian/sdb2/work/styled_caption/data/clf_cnn/word2idx.json'

在前几周我发现了您的论文并发现其中的idea对我的研究有大的帮助，但是我发现想复现这个文章的效果是困难的。在config.py里，存在大量不存在于本仓库的路径，这些路径往往指向一些资源文件，并在代码里面都起到了相当的作用，我想相关的资源也非常的重要，如果您能看见这个issue的话，方便您更新一下资源或者提供资源的获取方式么，谢谢。

Hi, wentian, I am very interested in MemCap. I would like to get the sent_sg_parse since use_sg is True for the train. But I failed to find where and how to generate based on the code. I am thinking it is the part called "Stylized Sentence Decomposing" in the paper. Any implementation on how to generate that?

Thank you for your excellent work and timely open source code. 👍
However, I don't know where to start running the code and configuring its dependencies. 😢
Can you provide more detail in the document? 💯
Thx a lot.