Downloads · 30 days
0
Physics-Inspired-Landscape-Learning-and-Flow-Matching
NYCU-MLLab/Physics-Inspired-Landscape-Learning-and-Flow-Matching
Physics-Inspired-Landscape-Learning-and-Flow-Matching is a other model from NYCU-MLLab. Use it for the other task on the model card, and read the license before you ship it in a product. It is set up for pytorch.
This repository stores the frozen PAFlow base checkpoint, its atom-number predictor, and the LandFlow route-potential checkpoint used in the main CrossDocked2020 experiment.
Access
Public
Updated Aug 24, 2026
Repo size
34.6 MB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.pt34.6 MB · 100%
At a glance
- Task
- other
- Library
- pytorch
- Access
- Public
- Created
- Aug 21, 2026
- Updated
- Aug 24, 2026
- SHA
- 6388ab44
Model card
From the Hugging Face model README
Physics-Inspired Landscape Learning and Flow Matching
LandFlow checkpoints
This repository stores the frozen PAFlow base checkpoint, its atom-number predictor, and the LandFlow route-potential checkpoint used in the main CrossDocked2020 experiment.
The LandFlow potential is a pairwise scalar model with 115,842 parameters. It predicts a trajectory-derived future-quality cost and supplies a normalized negative coordinate gradient during sampling. The PAFlow parameters remain frozen.
Use these files with the matching GitHub release. SHA-256 checksums are listed
in artifacts.json in that repository.
LandFlow is implemented on top of the PAFlow research code and includes frozen PAFlow checkpoints. The upstream repository did not contain a license when this release was prepared; verify the upstream redistribution terms before mirroring these files.
This model is for research use in molecular generation. Generated structures must not be interpreted as experimentally validated binders or medical advice.
- Task
- other
- Library
- pytorch
- Created
- Aug 21, 2026
- Updated
- Aug 24, 2026