
Loading weights:   0%|          | 0/311 [00:00<?, ?it/s]
Loading weights: 100%|██████████| 311/311 [00:00<00:00, 4414.59it/s]
full 0.0001 step 0 loss 1.34 exact 5/32 train_s 0
full 0.0001 step 4 loss 2.2514 exact 0/32 train_s 5.88
full 0.0001 step 8 loss 2.366 exact 0/32 train_s 11.67
full 0.0001 step 16 loss 2.2773 exact 0/32 train_s 23.23
Saved artifacts/qwen3-06b/2026-09-08-efficiency-screen/full-lr1e-4
