Bio Mega Server

Dashboard Workspace Tools Jobs

GPU backend: Boltz-2 submits through the Slurm/Singularity runner adapter. Other tools stay visible until their adapters are configured.

Back

Finetuning and Active Learning

Boltz-2 Affinity Finetuning

PaperCodeNeed >5GB

Finetune the Boltz-2 affinity head on your protein-ligand binding data

Details Paper Code

Job Name

Log in to submit jobs to the Slurm backend.

EXAMPLES

Training Data File *A3M FilesEpochsSamples per EpochLearning Rate

Tamarind API type: boltz-affinity-finetune.

Readiness

Readiness: Need >5GB
Smoke: not_tested
Artifacts: not_applicable
Production: not_started
Storage: >5
Strategy: large shared SIF/database/model bundle phase
Next: Defer to large-model/database phase or allow larger shared storage bundle.

large model weights, structure DBs, docking/MD assets, or GPU stack likely >5GB

NOS

Generate antibodies optimized for a given property

Finetuning and Active Learning

SaProt Finetuning

SaProt finetuning

Finetuning and Active Learning

Finetune Protein Language Model

Protein sequence property prediction finetuning

Finetuning and Active Learning

Finetune BALM

Protein-ligand binding affinity prediction finetuning

Finetuning and Active Learning

ProteusAI

ML-assisted Directed Evolution

Finetuning and Active Learning