EnzyGen2 Finetuning

PaperCodeNeed >5GB

Finetune EnzyGen2 on a specific enzyme family

Readiness

Readiness: Need >5GB
Smoke: not_tested
Artifacts: not_applicable
Production: need_gt5gb_cuda_fairseq_esm_finetune_pending
Storage: 5-20GB
Strategy: shared CUDA PyTorch/fairseq/ESM SIF + finetune data/model cache overlay
Next: After EnzyGen2 shared CUDA/fairseq/ESM SIF and checkpoints are staged, run a tiny bounded finetune Slurm smoke with enzyme-family JSON, then wire runner artifacts.

No obvious large model/database dependency identified; validate upstream docs before install. 2026-07-05 SOP batch recheck: backend inventory found no staged EnzyGen2 finetune source/checkpoints/data/SIF/artifacts/runner. Upstream finetuning depends on pretrained EnzyGen2 checkpoint, ESM2 650M, CUDA PyTorch/fairseq, and enzyme-family JSON data. Reclassified from Smoke Pending to Need >5GB.