These are chat archives for cltk/cltk

Aug 2017
Aug 16 2017 09:55
hi all, i've been working on ancient Chinese poems generation lately and i've got some work done, i believe these work fit cltk well, how could i make it into cltk?
the work which may help: ancient char and modern char into same vector space so they can calculate cosine similarities or find one modern char's most similar ancient chinese chars
how i do this: i use a pretrained w2v model and a modern chinese dict and ancient chinese dict, use the dictionary explanation to build the char vector via sent2vec
Kyle P. Johnson
Aug 16 2017 13:38
@PengFoo This sounds really interesting, but it may be too creative for the CLTK. We normally want to help researchers and students study ancient languages. We prefer to use techniques that have been accepted and published by scholars.
But we should understand more … would you share a link to the code, along with examples of its input and output?