-
step 280 epoch 0: maxsim_eval=1.0000 mrr_at_10=0.6390 pooled_mrr_at_10=0.4358 pooled_recall_at_100=0.9720 pooled_recall_at_20=0.7520 pooled_recall_at_5=0.5800 recall_at_100=0.9720 recall_at_20=0.9020 recall_at_5=0.8120
-
step 240 epoch 0: maxsim_eval=1.0000 mrr_at_10=0.6532 pooled_mrr_at_10=0.4390 pooled_recall_at_100=0.9720 pooled_recall_at_20=0.7300 pooled_recall_at_5=0.5800 recall_at_100=0.9720 recall_at_20=0.9340 recall_at_5=0.8660
-
step 200 epoch 0: maxsim_eval=1.0000 mrr_at_10=0.6655 pooled_mrr_at_10=0.5034 pooled_recall_at_100=0.9940 pooled_recall_at_20=0.8140 pooled_recall_at_5=0.6700 recall_at_100=0.9940 recall_at_20=0.9380 recall_at_5=0.8740
-
step 160 epoch 0: maxsim_eval=1.0000 mrr_at_10=0.4880 pooled_mrr_at_10=0.3941 pooled_recall_at_100=0.9460 pooled_recall_at_20=0.6400 pooled_recall_at_5=0.5340 recall_at_100=0.9460 recall_at_20=0.7020 recall_at_5=0.6240
-
step 120 epoch 0: maxsim_eval=1.0000 mrr_at_10=0.5663 pooled_mrr_at_10=0.4638 pooled_recall_at_100=0.9920 pooled_recall_at_20=0.7800 pooled_recall_at_5=0.6140 recall_at_100=0.9920 recall_at_20=0.8460 recall_at_5=0.7480
-
CTVA-101 Remove step_150 (trained 50 steps under mis-scaled temperature); resume anchors on step_100
-
step 150 epoch 0: maxsim_eval=1.0000 mrr_at_10=0.3538 pooled_mrr_at_10=0.4278 pooled_recall_at_100=0.9680 pooled_recall_at_20=0.7700 pooled_recall_at_5=0.6100 recall_at_100=0.9680 recall_at_20=0.6020 recall_at_5=0.4620
-
CTVA-101 Remove gate-runaway checkpoints step_150-step_450 so resume rescues from step_100 (g=0.049, MRR 0.4337)
-
step 450 epoch 0: maxsim_eval=1.0000 mrr_at_10=0.2456 pooled_mrr_at_10=0.5326 pooled_recall_at_100=0.9860 pooled_recall_at_20=0.8620 pooled_recall_at_5=0.7420 recall_at_100=0.9860 recall_at_20=0.4840 recall_at_5=0.3240
-
step 400 epoch 0: maxsim_eval=1.0000 mrr_at_10=0.2330 pooled_mrr_at_10=0.5049 pooled_recall_at_100=0.9700 pooled_recall_at_20=0.8420 pooled_recall_at_5=0.6760 recall_at_100=0.9700 recall_at_20=0.4520 recall_at_5=0.3280
-
step 350 epoch 0: maxsim_eval=1.0000 mrr_at_10=0.2573 pooled_mrr_at_10=0.4977 pooled_recall_at_100=0.9620 pooled_recall_at_20=0.7920 pooled_recall_at_5=0.6460 recall_at_100=0.9620 recall_at_20=0.4740 recall_at_5=0.3420
-
step 300 epoch 0: maxsim_eval=1.0000 mrr_at_10=0.2702 pooled_mrr_at_10=0.4529 pooled_recall_at_100=0.9640 pooled_recall_at_20=0.7380 pooled_recall_at_5=0.5940 recall_at_100=0.9640 recall_at_20=0.4920 recall_at_5=0.3500
-
step 250 epoch 0: maxsim_eval=1.0000 mrr_at_10=0.2658 pooled_mrr_at_10=0.4140 pooled_recall_at_100=0.9580 pooled_recall_at_20=0.7100 pooled_recall_at_5=0.5640 recall_at_100=0.9580 recall_at_20=0.5120 recall_at_5=0.3640
-
step 200 epoch 0: maxsim_eval=1.0000 mrr_at_10=0.3058 pooled_mrr_at_10=0.4013 pooled_recall_at_100=0.9220 pooled_recall_at_20=0.7160 pooled_recall_at_5=0.5480 recall_at_100=0.9220 recall_at_20=0.5660 recall_at_5=0.3980
-
step 150 epoch 0: maxsim_eval=1.0000 mrr_at_10=0.3436 pooled_mrr_at_10=0.4200 pooled_recall_at_100=0.9700 pooled_recall_at_20=0.7620 pooled_recall_at_5=0.6020 recall_at_100=0.9700 recall_at_20=0.5960 recall_at_5=0.4580
-
step 100 epoch 0: maxsim_eval=1.0000 mrr_at_10=0.4337 pooled_mrr_at_10=0.3940 pooled_recall_at_100=0.9460 pooled_recall_at_20=0.7120 pooled_recall_at_5=0.5160 recall_at_100=0.9460 recall_at_20=0.6980 recall_at_5=0.5580
-
step 50 epoch 0: maxsim_eval=1.0000 mrr_at_10=0.3119 pooled_mrr_at_10=0.2762 pooled_recall_at_100=0.8420 pooled_recall_at_20=0.5460 pooled_recall_at_5=0.3760 recall_at_100=0.8420 recall_at_20=0.5660 recall_at_5=0.4160
-
initial commit
