search

geek-ai / Texygen

A text generation benchmarking platform
MIT License
863 stars 203 forks source link

LeakGAN architecture #11

Closed pclucas14 closed 6 years ago

pclucas14 commented 6 years ago

Hi,

It seems that the default LeakGAN discriminator architecture in this repository is quite different from the one in the official repo. Specifically, Texygen's implementation has only two layers, while the offical's has over 10 layers. Is there any reason for this discrepency ?

Thanks in advance, Lucas

Yaoming95 commented 6 years ago

Sorry for the late reply. You can change the filter_size and num_filter. Here we use a smaller CNN compared to the original paper because of the limitation of computational resources. Since we use the same settings among all the baseline models (i.e. the filter is the same in SeqGAN and LeakGAN), we can still regard Texygen offered a relatively fair comparison.