@gaozhangyang
Hello
Thank you for developing the FoldToken-open project!
I have forked this project and added two lightweight new scripts:
Apart from simpler I/O and no batching, they follow the logic in your original reconstruct.py.
Could you please confirm whether these minimal pipelines are technically correct for converting PDBs to VQ-VAE IDs and vice versa, or point out any steps I should adjust?
Thanks