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

config: production epochs 100→200, min_epochs_before_stopping 50→100

Browse Source 
H100 has the compute budget — 200 epochs with cosine decay gives the
model full room to learn through C51 transition and converge.
Early stopping can't fire before epoch 100.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
This commit is contained in:
jgrusewski
2026-03-29 18:22:21 +02:00
parent 7be192a737
commit ae704a7b73
1 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
config/training/dqn-production.toml
View File

@@ -7,7 +7,7 @@
# - Cosine LR decay: 1e-4 → 1e-5 over 100 epochs # - Cosine LR decay: 1e-4 → 1e-5 over 100 epochs
[training] [training]
epochs = 100 epochs = 200
batch_size = 1024 batch_size = 1024
learning_rate = 0.0001 learning_rate = 0.0001
gamma = 0.99 gamma = 0.99
@@ -42,7 +42,7 @@ per_beta_start = 0.4
[early_stopping] [early_stopping]
enabled = true enabled = true
patience = 20 patience = 20
min_epochs_before_stopping = 50 min_epochs_before_stopping = 100
[experience] [experience]
initial_capital = 100000.0 initial_capital = 100000.0