GlycoBART Reproduction (207M)
Converged PyTorch checkpoint (best_model.pt, random initialization seed 0) for "GlycoGen: Glycan Structure Prediction with Crystallizing Flows" (NeurIPS 2026).
- Architecture / Objective: From-scratch reproduction of Abtheen et al. (2025) on CandyCrunch v2 benchmark
- Test-set top-1 structural accuracy (
n = 76,774):80.92% ± 0.31% (beam=1) - Code & inference scripts: github.com/rlacombe/GlycoGen
- Project page: rlacombe.github.io/GlycoGen
- Benchmark splits:
rlacombe/candycrunch-v2-benchmark
Citation
@inproceedings{lacombe2026glycogen,
title = {GlycoGen: Glycan Structure Prediction with Crystallizing Flows},
author = {Lacombe, Romain},
booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
year = {2026}
}