gitadmin/foxhunt
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

fix: train_walk_forward uses regime-stratified fold generation

Browse Source 
Was calling generate_folds() (non-stratified) instead of
generate_folds_stratified(). Walk-forward folds are now balanced
across trending/ranging/volatile regimes.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
This commit is contained in:
jgrusewski
2026-04-05 21:51:13 +02:00
parent 4956de7c74
commit f2b777a6ff
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
crates/ml/src/trainers/dqn/trainer/mod.rs
View File

@@ -567,7 +567,7 @@ impl DQNTrainer {
step_fraction: self.hyperparams.wf_step_fraction,
};
let folds = wf_config.generate_folds(features.len());
let folds = wf_config.generate_folds_stratified(training_data);
let num_folds = folds.len();
if num_folds == 0 {
return Err(anyhow::anyhow!(