DynamicsPLM Finetuning

PaperCodeNeed >5GB

Finetune a dynamics-aware protein language model on your data

Readiness

Readiness: Need >5GB
Smoke: not_tested
Artifacts: not_applicable
Production: need_gt5gb_saprot_lmdb_cuda_runtime_pending
Storage: 5-20GB
Strategy: shared CUDA PyTorch/SaProt SIF + LMDB/model cache overlay
Next: Build shared CUDA 11.6/PyTorch 1.13 SIF, stage SaProt_650M_AF2 weights and minimal LMDB dataset, run direct Slurm finetune smoke, then add runner adapter.

No obvious large model/database dependency identified; validate upstream docs before install. 2026-07-05 SOP batch recheck: backend inventory found no staged DynamicsPLM source/SaProt weights/LMDB data/SIF/artifacts/runner. Upstream requires Python 3.10 environment, LMDB datasets, and SaProt_650M_AF2 weights with model files around 2.6 GB each before runtime/data. Reclassified from Smoke Pending to Need >5GB.