Compiler Guided Generation Network (CGGNet)
Currently, I am working on cleaning up the code. It will be updated as soon as possible
Generated data evaluated in the evaluation section : https://1drv.ms/u/s!AmtjNBldUr6E1mj5VocGkZNPXub_?e=aWuPKM
4,043,399 unique and compilable files : https://1drv.ms/u/s!AmtjNBldUr6E1mmkXCXeLbjRhInv?e=nkarQl
==============================================================
I am currently in the process of cleaning up the relevant code. However, due to personal issues and environmental constraints, I am unable to update it immediately. I apologize for this delay and appreciate your understanding. For now, I am attaching the source code that can be used in the paper. You can obtain results by running run_seqgan.py (although it's currently named seqGAN, it's actually CGGNet, a modified version of SeqGAN - it's not SeqGAN!). At present, there are numerous unnecessary dependency files and code segments. We plan to modify these relevant parts as quickly as possible.
Thanks and Refer to: https://github.com/williamSYSU/TextGAN-PyTorch