|
Download README.md from MurrellLab/LPGM: direct link, hf CLI and curl.
- Browser
- Download file 2.08 kB
-
https://huggingface.co/MurrellLab/LPGM/resolve/main/README.md
- Command line
-
hf download hf://MurrellLab/LPGM/README.md
-
curl -L -o README.md https://huggingface.co/MurrellLab/LPGM/resolve/main/README.md
2.08 kB
| library_name: pytorch | |
| tags: | |
| - molecular-generation | |
| - flow-matching | |
| - geom | |
| - qm9 | |
| # Molecular LPGM | |
| [Reference implementation and documentation](https://github.com/MurrellGroup/molLPGM) | |
| Final **Fragment Noise** and **Brownian Atoms** models for GEOM-Drugs and QM9. | |
| Both methods condition on atom count and the BRICS fragment partition, with | |
| learned same/different-fragment attention biases in every transformer layer. | |
| Neither receives clean coordinates, atom labels, bonds or geometric templates | |
| at generation. Fragment Noise uses a single fixed global angular variance of | |
| **6** in both datasets. | |
| | Dataset | Width | Layers | Fragment Noise | Brownian Atoms | | |
| | --- | ---: | ---: | --- | --- | | |
| | GEOM-Drugs | 768 | 16 | [fragment.pt](geom/fragment.pt) | [bm.pt](geom/bm.pt) | | |
| | QM9 | 512 | 12 | [fragment.pt](qm9/fragment.pt) | [bm.pt](qm9/bm.pt) | | |
| All four checkpoints contain final **800,000-update raw and EMA weights** | |
| (decay **0.9999**), architecture, process parameters, and final loss weights. | |
| The final 100,000 updates are learning-rate warmdown. Optimizer state is omitted. | |
| Use `--ema` in the reference sampling commands, or `load_checkpoint(..., | |
| use_ema=True)` in Python. SHA-256 hashes and parameter counts are in | |
| [checkpoints.json](checkpoints.json). | |
| The source is a centered Gaussian with standard deviation 3 Å. Brownian Atoms | |
| uses the scalar clock `u(t) = 1 - (1 - t**2.3)**3.1` and variance coefficient | |
| `beta = 14.8 Ų`; these parameters approximately match the conditional | |
| endpoint spread. It has one Cartesian coordinate head and the same atom-label | |
| process as Fragment Noise. | |
| [10K-sample results and archives](https://github.com/MurrellGroup/molLPGM/tree/main/results) | |
| use final EMA weights and 2,000 square-root-spaced steps. The same ADiT validity, | |
| valid-and-connected, and uniqueness protocol is used for both datasets. | |
| GEOM evaluation uses test cases; QM9 pools validation and test cases. | |
| See the [comparison protocol](https://github.com/MurrellGroup/molLPGM/blob/main/docs/comparison.md) | |
| for conditioning, objective and training details. | |