boltz issues

Trunk sampling for increased diversity

5

Hi Boltz team, Is it possible to increase diversity in boltz-1 predictions by doing multiple trunk sampling iterations (each time with a different seed), like what AlphaFold3 reported in their...

rui-yin

Checking similarity of input ligand w.r.t training set

As described in the preprint, one of the criteria for constructing the validation set is "at least one of the small-molecules exhibits a tanimoto similarity of 0.8 or less to...

zhang-ivy

RTX 50XX (sm120) vs Pytoch/cuda

Hi, are you planning to update Boltz1 to support sm120 GPU (RTX 50XX) so the latest Pytorch and Cuda (12.8)? In the dependencies you have (torch>=2.2) pytorch-lightning==2.4.0 fairscale==0.4.13 These would...

CLG68

Chirality of Ligand Not Preserved

When I predict the structures of protein-ligand complexes using Boltz, the chirality of the input ligands is not being preserved. Some of the ligands have a chiral phosphorus atom and...

rjrich

Issue of failed featurizer from version 0.4.1

2

Hi, I was using Boltz-1 version 0.4.1 for some inference tasks, among which some failed with a series of errors like below and it took a long time to fail:...

weitse-hsu

retraining the model

1

Hi, Can I retrain the model our customized or proprietary dataset ?

wangjiawen2013

pair_chains_iptm score returned when folding a single chain

1

When I run Boltz with a single protein subunit (no multiple chains) and a ligand I still get a pair_chains_iptm score and other scores associated with chain interaction. Why is...

avaektefaie