Progress stuck training forests

[vody-am]

Hi!

I've been testing this application out on Delicious-200K from the XML Repository, and it appears I've gotten stuck at 90.21%.
I ran with the following parameters:

+ omikuji train --model_path ./deli_model --cluster.k 100 --max_depth 3 --cluster.unbalanced deliciousLarge_train.txt
2025-05-31T17:21:43.666Z INFO  [omikuji::data] Loading data from deliciousLarge_train.txt
2025-05-31T17:21:43.992Z INFO  [omikuji::data] Parsing data
2025-05-31T17:21:44.370Z INFO  [omikuji::data] Loaded 196606 examples; it took 0.70s
2025-05-31T17:21:44.383Z INFO  [omikuji::model::train] Training model with hyper-parameters HyperParam { n_trees: 3, min_branch_size: 100, max_depth: 3, centroid_threshold: 0.0, collapse_every_n_layers: 0, linear: HyperParam { loss_type: Hinge, eps: 0.1, c: 1.0, weight_threshold: 0.1, max_iter: 20 }, cluster: HyperParam { k: 100, balanced: false, eps: 0.0001, min_size: 2 }, tree_structure_only: false, train_trees_1_by_1: false }
2025-05-31T17:21:44.383Z INFO  [omikuji::model::train] Initializing tree trainer
2025-05-31T17:21:44.397Z INFO  [omikuji::model::train] Computing label centroids
Labels 201353 / 201353 [=============================================================================================================] 100.00 % 3459.92/s 2025-05-31T17:33:06.831Z INFO  [omikuji::model::train] Start training forest
560405 / 621194 [=======================================================================================================>-----------] 90.21 % 20.12/s 50m 

not quite sure what's going on there, but it appears stuck at 100% CPU util on all cores, without progressing. Possibly some kind of threading bug?

I'm on Apple M4 CPU if that helps at all. I'll do a few re-runs to see if I can reproduce.