Search results

Add lightweight Bahdanau-style attention pooling for site aggregation

Status: Open.

#45 In matsengrp/dpvt;

· matsen opened on May 7, 2026

Dynamic per-sample class reweighting for BCE loss

Status: Open.

#43 In matsengrp/dpvt;

· lenacoll opened on Feb 26, 2026

Log hyperparameters when training

Status: Open.

#31 In matsengrp/dpvt;

· matsen opened on Nov 14, 2024

Improve NN models by using batch-wise operations

Status: Open.

#24 In matsengrp/dpvt;

· harryrichman opened on Apr 18, 2024

Separate training of traversal-step NN and site-aggregration-step NN

Status: Open.

#17 In matsengrp/dpvt;

· harryrichman opened on Mar 30, 2024

How are we going to scale up training and testing?

Status: Open.

#15 In matsengrp/dpvt;

· matsen opened on Mar 2, 2024

Could we characterize the hardness of this problem?

Status: Open.

#12 In matsengrp/dpvt;

· matsen opened on Feb 25, 2024

Implement multiprocessing in forward function

Status: Open.

#9 In matsengrp/dpvt;

· harryrichman opened on Feb 20, 2024

perfect phylogenies should minimize parsimony

Status: Open.

#6 In matsengrp/dpvt;

· chrisjenningsshaffer opened on Feb 7, 2024

Perturb trees, then run Sankoff

Status: Open.

#3 In matsengrp/dpvt;

· matsen opened on Jan 24, 2024