search

tech-srl / code2seq

Code for the model presented in the paper: "code2seq: Generating Sequences from Structured Representations of Code"
http://code2seq.org
MIT License
548 stars 165 forks source link

Fine-tuning tips #97

Closed DRMALEK closed 3 years ago

DRMALEK commented 3 years ago

Hi,

I'm trying to fine-tune the code2seq model that is trained on the deepcom dataset on another more specific dataset (Android third-party libraries), so what are the general tips or guidelines for achieving such a thing.

Thanks in advance.

urialon commented 3 years ago

Hi @DRMALEK , I don't have any specific tips or guidelines, but this sounds like a very interesting direction. Feel free to share the results or PDF.

Best, Uri