Archive 4 checkpoints (1-4)

paper-releases/pbisc-diffusion/updates/20261003T103040Z-crosscohort-k1000-saturation/diffusion/dataset/server_1/model8/plans/paired_crosscohort_20261002/runs/hoL_immicro/K1000/lora64sat/train.log

@@ -0,0 +1,161 @@
+2026-10-03T07:56:07 split=hoL_immicro K=1000 mode=lora trainable=5,184,896 balance=cohort ye2026_sepsis: train 73 rows / 529,816 cells, inner 9; imperial_eskd: train 146 rows / 462,172 cells, inner 11; giroux2022: train 24 rows / 254,250 cells, inner 0; liu2021: train 40 rows / 147,532 cells, inner 4
+2026-10-03T07:56:08 step 0 inner-val loss 0.11485 (pretrained) {'ye2026_sepsis': 0.12609728963838684, 'imperial_eskd': 0.1086193590678952, 'liu2021': 0.10984707903116941}
+2026-10-03T07:56:19 step 250 train 0.10536 inner-val 0.11383 best 0.11383@250 (11s) ye2026_sepsis=0.1251 imperial_eskd=0.1075 liu2021=0.1089
+2026-10-03T07:56:31 step 500 train 0.10306 inner-val 0.11182 best 0.11182@500 (23s) ye2026_sepsis=0.1230 imperial_eskd=0.1053 liu2021=0.1071
+2026-10-03T07:56:43 step 750 train 0.10288 inner-val 0.10965 best 0.10965@750 (35s) ye2026_sepsis=0.1205 imperial_eskd=0.1031 liu2021=0.1054
+2026-10-03T07:56:55 step 1000 train 0.10245 inner-val 0.10803 best 0.10803@1000 (47s) ye2026_sepsis=0.1184 imperial_eskd=0.1015 liu2021=0.1042
+2026-10-03T07:57:06 step 1250 train 0.10268 inner-val 0.10708 best 0.10708@1250 (58s) ye2026_sepsis=0.1171 imperial_eskd=0.1006 liu2021=0.1035
+2026-10-03T07:57:16 step 1500 train 0.10254 inner-val 0.10653 best 0.10653@1500 (68s) ye2026_sepsis=0.1164 imperial_eskd=0.1000 liu2021=0.1031
+2026-10-03T07:57:26 step 1750 train 0.10235 inner-val 0.10617 best 0.10617@1750 (78s) ye2026_sepsis=0.1160 imperial_eskd=0.0997 liu2021=0.1029
+2026-10-03T07:57:37 step 2000 train 0.10219 inner-val 0.10594 best 0.10594@2000 (88s) ye2026_sepsis=0.1157 imperial_eskd=0.0994 liu2021=0.1027
+2026-10-03T07:57:47 step 2250 train 0.10219 inner-val 0.10578 best 0.10578@2250 (99s) ye2026_sepsis=0.1155 imperial_eskd=0.0993 liu2021=0.1026
+2026-10-03T07:57:57 step 2500 train 0.10270 inner-val 0.10567 best 0.10567@2500 (109s) ye2026_sepsis=0.1154 imperial_eskd=0.0992 liu2021=0.1025
+2026-10-03T07:58:08 step 2750 train 0.10217 inner-val 0.10559 best 0.10559@2750 (120s) ye2026_sepsis=0.1153 imperial_eskd=0.0991 liu2021=0.1024
+2026-10-03T07:58:18 step 3000 train 0.10217 inner-val 0.10553 best 0.10553@3000 (130s) ye2026_sepsis=0.1152 imperial_eskd=0.0991 liu2021=0.1023
+2026-10-03T07:58:29 step 3250 train 0.10250 inner-val 0.10550 best 0.10550@3250 (140s) ye2026_sepsis=0.1151 imperial_eskd=0.0990 liu2021=0.1023
+2026-10-03T07:58:39 step 3500 train 0.10241 inner-val 0.10546 best 0.10546@3500 (151s) ye2026_sepsis=0.1151 imperial_eskd=0.0990 liu2021=0.1023
+2026-10-03T07:58:49 step 3750 train 0.10206 inner-val 0.10544 best 0.10544@3750 (161s) ye2026_sepsis=0.1151 imperial_eskd=0.0990 liu2021=0.1023
+2026-10-03T07:59:00 step 4000 train 0.10216 inner-val 0.10542 best 0.10542@4000 (171s) ye2026_sepsis=0.1151 imperial_eskd=0.0989 liu2021=0.1023
+2026-10-03T07:59:10 step 4250 train 0.10214 inner-val 0.10541 best 0.10541@4250 (182s) ye2026_sepsis=0.1150 imperial_eskd=0.0989 liu2021=0.1023
+2026-10-03T07:59:20 step 4500 train 0.10216 inner-val 0.10539 best 0.10539@4500 (192s) ye2026_sepsis=0.1150 imperial_eskd=0.0989 liu2021=0.1022
+2026-10-03T07:59:30 step 4750 train 0.10217 inner-val 0.10538 best 0.10538@4750 (202s) ye2026_sepsis=0.1150 imperial_eskd=0.0989 liu2021=0.1022
+2026-10-03T07:59:41 step 5000 train 0.10214 inner-val 0.10537 best 0.10537@5000 (212s) ye2026_sepsis=0.1150 imperial_eskd=0.0989 liu2021=0.1022
+2026-10-03T07:59:51 step 5250 train 0.10207 inner-val 0.10536 best 0.10536@5250 (223s) ye2026_sepsis=0.1150 imperial_eskd=0.0988 liu2021=0.1022
+2026-10-03T08:00:01 step 5500 train 0.10211 inner-val 0.10535 best 0.10535@5500 (233s) ye2026_sepsis=0.1150 imperial_eskd=0.0988 liu2021=0.1022
+2026-10-03T08:00:12 step 5750 train 0.10171 inner-val 0.10534 best 0.10534@5750 (244s) ye2026_sepsis=0.1150 imperial_eskd=0.0988 liu2021=0.1022
+2026-10-03T08:00:22 step 6000 train 0.10194 inner-val 0.10533 best 0.10533@6000 (254s) ye2026_sepsis=0.1150 imperial_eskd=0.0988 liu2021=0.1022
+2026-10-03T08:00:32 step 6250 train 0.10207 inner-val 0.10532 best 0.10532@6250 (264s) ye2026_sepsis=0.1150 imperial_eskd=0.0988 liu2021=0.1022
+2026-10-03T08:00:43 step 6500 train 0.10209 inner-val 0.10532 best 0.10532@6500 (274s) ye2026_sepsis=0.1150 imperial_eskd=0.0988 liu2021=0.1022
+2026-10-03T08:00:53 step 6750 train 0.10201 inner-val 0.10531 best 0.10531@6750 (285s) ye2026_sepsis=0.1150 imperial_eskd=0.0988 liu2021=0.1022
+2026-10-03T08:01:03 step 7000 train 0.10182 inner-val 0.10530 best 0.10530@7000 (295s) ye2026_sepsis=0.1150 imperial_eskd=0.0988 liu2021=0.1021
+2026-10-03T08:01:14 step 7250 train 0.10208 inner-val 0.10529 best 0.10529@7250 (305s) ye2026_sepsis=0.1150 imperial_eskd=0.0987 liu2021=0.1021
+2026-10-03T08:01:24 step 7500 train 0.10195 inner-val 0.10529 best 0.10529@7500 (316s) ye2026_sepsis=0.1150 imperial_eskd=0.0987 liu2021=0.1021
+2026-10-03T08:01:34 step 7750 train 0.10192 inner-val 0.10528 best 0.10528@7750 (326s) ye2026_sepsis=0.1150 imperial_eskd=0.0987 liu2021=0.1021
+2026-10-03T08:01:45 step 8000 train 0.10155 inner-val 0.10528 best 0.10528@8000 (336s) ye2026_sepsis=0.1150 imperial_eskd=0.0987 liu2021=0.1021
+2026-10-03T08:01:55 step 8250 train 0.10192 inner-val 0.10528 best 0.10528@8250 (347s) ye2026_sepsis=0.1150 imperial_eskd=0.0987 liu2021=0.1021
+2026-10-03T08:02:05 step 8500 train 0.10162 inner-val 0.10527 best 0.10527@8500 (357s) ye2026_sepsis=0.1150 imperial_eskd=0.0987 liu2021=0.1021
+2026-10-03T08:02:16 step 8750 train 0.10223 inner-val 0.10527 best 0.10527@8750 (368s) ye2026_sepsis=0.1150 imperial_eskd=0.0987 liu2021=0.1021
+2026-10-03T08:02:26 step 9000 train 0.10202 inner-val 0.10526 best 0.10526@9000 (378s) ye2026_sepsis=0.1150 imperial_eskd=0.0987 liu2021=0.1021
+2026-10-03T08:02:36 step 9250 train 0.10186 inner-val 0.10525 best 0.10525@9250 (388s) ye2026_sepsis=0.1150 imperial_eskd=0.0987 liu2021=0.1021
+2026-10-03T08:02:47 step 9500 train 0.10187 inner-val 0.10525 best 0.10525@9250 (398s) ye2026_sepsis=0.1150 imperial_eskd=0.0987 liu2021=0.1021
+2026-10-03T08:02:57 step 9750 train 0.10189 inner-val 0.10526 best 0.10525@9250 (409s) ye2026_sepsis=0.1150 imperial_eskd=0.0987 liu2021=0.1021
+2026-10-03T08:03:07 step 10000 train 0.10184 inner-val 0.10525 best 0.10525@10000 (419s) ye2026_sepsis=0.1150 imperial_eskd=0.0987 liu2021=0.1021
+2026-10-03T08:03:18 step 10250 train 0.10191 inner-val 0.10525 best 0.10525@10250 (429s) ye2026_sepsis=0.1150 imperial_eskd=0.0987 liu2021=0.1021
+2026-10-03T08:03:28 step 10500 train 0.10207 inner-val 0.10524 best 0.10524@10500 (440s) ye2026_sepsis=0.1150 imperial_eskd=0.0987 liu2021=0.1021
+2026-10-03T08:03:38 step 10750 train 0.10194 inner-val 0.10524 best 0.10524@10750 (450s) ye2026_sepsis=0.1149 imperial_eskd=0.0987 liu2021=0.1021
+2026-10-03T08:03:48 step 11000 train 0.10174 inner-val 0.10523 best 0.10523@11000 (460s) ye2026_sepsis=0.1149 imperial_eskd=0.0987 liu2021=0.1021
+2026-10-03T08:03:59 step 11250 train 0.10173 inner-val 0.10523 best 0.10523@11250 (470s) ye2026_sepsis=0.1149 imperial_eskd=0.0987 liu2021=0.1021
+2026-10-03T08:04:09 step 11500 train 0.10175 inner-val 0.10522 best 0.10522@11500 (481s) ye2026_sepsis=0.1149 imperial_eskd=0.0987 liu2021=0.1020
+2026-10-03T08:04:19 step 11750 train 0.10172 inner-val 0.10522 best 0.10522@11750 (491s) ye2026_sepsis=0.1149 imperial_eskd=0.0987 liu2021=0.1020
+2026-10-03T08:04:30 step 12000 train 0.10153 inner-val 0.10522 best 0.10522@12000 (502s) ye2026_sepsis=0.1149 imperial_eskd=0.0987 liu2021=0.1020
+2026-10-03T08:04:40 step 12250 train 0.10171 inner-val 0.10521 best 0.10521@12250 (512s) ye2026_sepsis=0.1149 imperial_eskd=0.0987 liu2021=0.1020
+2026-10-03T08:04:50 step 12500 train 0.10190 inner-val 0.10521 best 0.10521@12500 (522s) ye2026_sepsis=0.1149 imperial_eskd=0.0987 liu2021=0.1020
+2026-10-03T08:05:00 step 12750 train 0.10167 inner-val 0.10520 best 0.10520@12750 (532s) ye2026_sepsis=0.1149 imperial_eskd=0.0987 liu2021=0.1020
+2026-10-03T08:05:11 step 13000 train 0.10191 inner-val 0.10520 best 0.10520@13000 (543s) ye2026_sepsis=0.1149 imperial_eskd=0.0987 liu2021=0.1020
+2026-10-03T08:05:21 step 13250 train 0.10186 inner-val 0.10520 best 0.10520@13250 (553s) ye2026_sepsis=0.1149 imperial_eskd=0.0987 liu2021=0.1020
+2026-10-03T08:05:32 step 13500 train 0.10164 inner-val 0.10519 best 0.10519@13500 (563s) ye2026_sepsis=0.1149 imperial_eskd=0.0986 liu2021=0.1020
+2026-10-03T08:05:42 step 13750 train 0.10185 inner-val 0.10518 best 0.10518@13750 (574s) ye2026_sepsis=0.1149 imperial_eskd=0.0986 liu2021=0.1020
+2026-10-03T08:05:52 step 14000 train 0.10176 inner-val 0.10517 best 0.10517@14000 (584s) ye2026_sepsis=0.1149 imperial_eskd=0.0986 liu2021=0.1020
+2026-10-03T08:06:02 step 14250 train 0.10138 inner-val 0.10517 best 0.10517@14250 (594s) ye2026_sepsis=0.1149 imperial_eskd=0.0986 liu2021=0.1020
+2026-10-03T08:06:13 step 14500 train 0.10170 inner-val 0.10516 best 0.10516@14500 (605s) ye2026_sepsis=0.1149 imperial_eskd=0.0986 liu2021=0.1020
+2026-10-03T08:06:23 step 14750 train 0.10199 inner-val 0.10516 best 0.10516@14750 (615s) ye2026_sepsis=0.1149 imperial_eskd=0.0986 liu2021=0.1020
+2026-10-03T08:06:33 step 15000 train 0.10192 inner-val 0.10515 best 0.10515@15000 (625s) ye2026_sepsis=0.1149 imperial_eskd=0.0986 liu2021=0.1020
+2026-10-03T08:06:44 step 15250 train 0.10124 inner-val 0.10515 best 0.10515@15000 (636s) ye2026_sepsis=0.1149 imperial_eskd=0.0986 liu2021=0.1020
+2026-10-03T08:06:54 step 15500 train 0.10184 inner-val 0.10515 best 0.10515@15500 (646s) ye2026_sepsis=0.1149 imperial_eskd=0.0986 liu2021=0.1020
+2026-10-03T08:07:05 step 15750 train 0.10158 inner-val 0.10515 best 0.10515@15750 (657s) ye2026_sepsis=0.1149 imperial_eskd=0.0986 liu2021=0.1020
+2026-10-03T08:07:15 step 16000 train 0.10153 inner-val 0.10514 best 0.10514@16000 (667s) ye2026_sepsis=0.1149 imperial_eskd=0.0986 liu2021=0.1020
+2026-10-03T08:07:25 step 16250 train 0.10183 inner-val 0.10514 best 0.10514@16000 (677s) ye2026_sepsis=0.1149 imperial_eskd=0.0986 liu2021=0.1020
+2026-10-03T08:07:36 step 16500 train 0.10174 inner-val 0.10514 best 0.10514@16500 (687s) ye2026_sepsis=0.1149 imperial_eskd=0.0986 liu2021=0.1020
+2026-10-03T08:07:46 step 16750 train 0.10163 inner-val 0.10514 best 0.10514@16750 (698s) ye2026_sepsis=0.1149 imperial_eskd=0.0986 liu2021=0.1020
+2026-10-03T08:07:56 step 17000 train 0.10192 inner-val 0.10513 best 0.10513@17000 (708s) ye2026_sepsis=0.1148 imperial_eskd=0.0986 liu2021=0.1020
+2026-10-03T08:08:06 step 17250 train 0.10189 inner-val 0.10513 best 0.10513@17250 (718s) ye2026_sepsis=0.1148 imperial_eskd=0.0986 liu2021=0.1019
+2026-10-03T08:08:17 step 17500 train 0.10198 inner-val 0.10513 best 0.10513@17250 (729s) ye2026_sepsis=0.1148 imperial_eskd=0.0986 liu2021=0.1019
+2026-10-03T08:08:27 step 17750 train 0.10137 inner-val 0.10512 best 0.10512@17750 (739s) ye2026_sepsis=0.1148 imperial_eskd=0.0986 liu2021=0.1019
+2026-10-03T08:08:38 step 18000 train 0.10204 inner-val 0.10512 best 0.10512@18000 (750s) ye2026_sepsis=0.1148 imperial_eskd=0.0986 liu2021=0.1019
+2026-10-03T08:08:48 step 18250 train 0.10146 inner-val 0.10511 best 0.10511@18250 (760s) ye2026_sepsis=0.1148 imperial_eskd=0.0986 liu2021=0.1019
+2026-10-03T08:08:58 step 18500 train 0.10178 inner-val 0.10511 best 0.10511@18500 (770s) ye2026_sepsis=0.1148 imperial_eskd=0.0986 liu2021=0.1019
+2026-10-03T08:09:09 step 18750 train 0.10147 inner-val 0.10510 best 0.10510@18750 (780s) ye2026_sepsis=0.1148 imperial_eskd=0.0986 liu2021=0.1019
+2026-10-03T08:09:19 step 19000 train 0.10163 inner-val 0.10510 best 0.10510@19000 (791s) ye2026_sepsis=0.1148 imperial_eskd=0.0986 liu2021=0.1019
+2026-10-03T08:09:29 step 19250 train 0.10152 inner-val 0.10509 best 0.10509@19250 (801s) ye2026_sepsis=0.1148 imperial_eskd=0.0986 liu2021=0.1019
+2026-10-03T08:09:39 step 19500 train 0.10165 inner-val 0.10509 best 0.10509@19500 (811s) ye2026_sepsis=0.1148 imperial_eskd=0.0986 liu2021=0.1019
+2026-10-03T08:09:50 step 19750 train 0.10152 inner-val 0.10508 best 0.10508@19750 (822s) ye2026_sepsis=0.1148 imperial_eskd=0.0986 liu2021=0.1019
+2026-10-03T08:10:00 step 20000 train 0.10150 inner-val 0.10508 best 0.10508@20000 (832s) ye2026_sepsis=0.1148 imperial_eskd=0.0986 liu2021=0.1019
+2026-10-03T08:10:02 block end 20000: best 0.10508@20000; block improvement 0.009776 (100.00 % of total 0.009776) -> continue
+2026-10-03T08:10:12 step 20250 train 0.10164 inner-val 0.10508 best 0.10508@20250 (844s) ye2026_sepsis=0.1148 imperial_eskd=0.0986 liu2021=0.1019
+2026-10-03T08:10:23 step 20500 train 0.10142 inner-val 0.10507 best 0.10507@20500 (855s) ye2026_sepsis=0.1148 imperial_eskd=0.0986 liu2021=0.1019
+2026-10-03T08:10:33 step 20750 train 0.10153 inner-val 0.10507 best 0.10507@20750 (865s) ye2026_sepsis=0.1148 imperial_eskd=0.0986 liu2021=0.1019
+2026-10-03T08:10:43 step 21000 train 0.10175 inner-val 0.10506 best 0.10506@21000 (875s) ye2026_sepsis=0.1148 imperial_eskd=0.0986 liu2021=0.1019
+2026-10-03T08:10:54 step 21250 train 0.10129 inner-val 0.10506 best 0.10506@21250 (886s) ye2026_sepsis=0.1148 imperial_eskd=0.0985 liu2021=0.1019
+2026-10-03T08:11:04 step 21500 train 0.10158 inner-val 0.10505 best 0.10505@21500 (896s) ye2026_sepsis=0.1148 imperial_eskd=0.0985 liu2021=0.1019
+2026-10-03T08:11:15 step 21750 train 0.10159 inner-val 0.10505 best 0.10505@21750 (906s) ye2026_sepsis=0.1148 imperial_eskd=0.0985 liu2021=0.1019
+2026-10-03T08:11:25 step 22000 train 0.10153 inner-val 0.10504 best 0.10504@22000 (917s) ye2026_sepsis=0.1148 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:11:35 step 22250 train 0.10165 inner-val 0.10504 best 0.10504@22250 (927s) ye2026_sepsis=0.1147 imperial_eskd=0.0985 liu2021=0.1019
+2026-10-03T08:11:46 step 22500 train 0.10148 inner-val 0.10504 best 0.10504@22500 (937s) ye2026_sepsis=0.1147 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:11:56 step 22750 train 0.10144 inner-val 0.10503 best 0.10503@22750 (948s) ye2026_sepsis=0.1147 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:12:06 step 23000 train 0.10189 inner-val 0.10503 best 0.10503@23000 (958s) ye2026_sepsis=0.1147 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:12:17 step 23250 train 0.10180 inner-val 0.10503 best 0.10503@23250 (968s) ye2026_sepsis=0.1147 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:12:27 step 23500 train 0.10162 inner-val 0.10503 best 0.10503@23500 (979s) ye2026_sepsis=0.1147 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:12:37 step 23750 train 0.10143 inner-val 0.10502 best 0.10502@23750 (989s) ye2026_sepsis=0.1147 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:12:47 step 24000 train 0.10195 inner-val 0.10502 best 0.10502@24000 (999s) ye2026_sepsis=0.1147 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:12:57 step 24250 train 0.10137 inner-val 0.10501 best 0.10501@24250 (1009s) ye2026_sepsis=0.1147 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:13:08 step 24500 train 0.10142 inner-val 0.10500 best 0.10500@24500 (1020s) ye2026_sepsis=0.1147 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:13:18 step 24750 train 0.10158 inner-val 0.10500 best 0.10500@24750 (1030s) ye2026_sepsis=0.1147 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:13:28 step 25000 train 0.10143 inner-val 0.10500 best 0.10500@25000 (1040s) ye2026_sepsis=0.1147 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:13:39 step 25250 train 0.10152 inner-val 0.10500 best 0.10500@25000 (1051s) ye2026_sepsis=0.1147 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:13:49 step 25500 train 0.10129 inner-val 0.10500 best 0.10500@25500 (1061s) ye2026_sepsis=0.1147 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:13:59 step 25750 train 0.10152 inner-val 0.10499 best 0.10499@25750 (1071s) ye2026_sepsis=0.1147 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:14:10 step 26000 train 0.10166 inner-val 0.10499 best 0.10499@26000 (1081s) ye2026_sepsis=0.1147 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:14:20 step 26250 train 0.10150 inner-val 0.10498 best 0.10498@26250 (1092s) ye2026_sepsis=0.1147 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:14:30 step 26500 train 0.10155 inner-val 0.10497 best 0.10497@26500 (1102s) ye2026_sepsis=0.1146 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:14:41 step 26750 train 0.10105 inner-val 0.10497 best 0.10497@26750 (1112s) ye2026_sepsis=0.1146 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:14:51 step 27000 train 0.10156 inner-val 0.10497 best 0.10497@27000 (1123s) ye2026_sepsis=0.1146 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:15:01 step 27250 train 0.10157 inner-val 0.10497 best 0.10497@27250 (1133s) ye2026_sepsis=0.1146 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:15:11 step 27500 train 0.10169 inner-val 0.10497 best 0.10497@27250 (1143s) ye2026_sepsis=0.1146 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:15:22 step 27750 train 0.10169 inner-val 0.10497 best 0.10497@27250 (1154s) ye2026_sepsis=0.1146 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:15:32 step 28000 train 0.10141 inner-val 0.10497 best 0.10497@28000 (1164s) ye2026_sepsis=0.1146 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:15:42 step 28250 train 0.10141 inner-val 0.10496 best 0.10496@28250 (1174s) ye2026_sepsis=0.1146 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:15:53 step 28500 train 0.10130 inner-val 0.10496 best 0.10496@28500 (1185s) ye2026_sepsis=0.1146 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:16:03 step 28750 train 0.10141 inner-val 0.10495 best 0.10495@28750 (1195s) ye2026_sepsis=0.1146 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:16:13 step 29000 train 0.10134 inner-val 0.10495 best 0.10495@29000 (1205s) ye2026_sepsis=0.1146 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:16:24 step 29250 train 0.10182 inner-val 0.10495 best 0.10495@29250 (1216s) ye2026_sepsis=0.1146 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:16:34 step 29500 train 0.10114 inner-val 0.10495 best 0.10495@29250 (1226s) ye2026_sepsis=0.1146 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:16:45 step 29750 train 0.10135 inner-val 0.10494 best 0.10494@29750 (1237s) ye2026_sepsis=0.1146 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:16:55 step 30000 train 0.10102 inner-val 0.10494 best 0.10494@29750 (1247s) ye2026_sepsis=0.1146 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:17:06 step 30250 train 0.10125 inner-val 0.10494 best 0.10494@30250 (1257s) ye2026_sepsis=0.1146 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:17:16 step 30500 train 0.10132 inner-val 0.10494 best 0.10494@30250 (1268s) ye2026_sepsis=0.1146 imperial_eskd=0.0984 liu2021=0.1018
+2026-10-03T08:17:26 step 30750 train 0.10127 inner-val 0.10494 best 0.10494@30250 (1278s) ye2026_sepsis=0.1146 imperial_eskd=0.0984 liu2021=0.1018
+2026-10-03T08:17:37 step 31000 train 0.10122 inner-val 0.10494 best 0.10494@31000 (1288s) ye2026_sepsis=0.1146 imperial_eskd=0.0985 liu2021=0.1018
+2026-10-03T08:17:47 step 31250 train 0.10126 inner-val 0.10494 best 0.10494@31250 (1299s) ye2026_sepsis=0.1146 imperial_eskd=0.0984 liu2021=0.1018
+2026-10-03T08:17:57 step 31500 train 0.10125 inner-val 0.10493 best 0.10493@31500 (1309s) ye2026_sepsis=0.1146 imperial_eskd=0.0984 liu2021=0.1018
+2026-10-03T08:18:08 step 31750 train 0.10123 inner-val 0.10492 best 0.10492@31750 (1319s) ye2026_sepsis=0.1146 imperial_eskd=0.0984 liu2021=0.1017
+2026-10-03T08:18:18 step 32000 train 0.10129 inner-val 0.10492 best 0.10492@32000 (1330s) ye2026_sepsis=0.1146 imperial_eskd=0.0984 liu2021=0.1017
+2026-10-03T08:18:28 step 32250 train 0.10130 inner-val 0.10492 best 0.10492@32250 (1340s) ye2026_sepsis=0.1146 imperial_eskd=0.0984 liu2021=0.1017
+2026-10-03T08:18:39 step 32500 train 0.10138 inner-val 0.10491 best 0.10491@32500 (1350s) ye2026_sepsis=0.1146 imperial_eskd=0.0984 liu2021=0.1017
+2026-10-03T08:18:49 step 32750 train 0.10131 inner-val 0.10491 best 0.10491@32750 (1361s) ye2026_sepsis=0.1146 imperial_eskd=0.0984 liu2021=0.1017
+2026-10-03T08:18:59 step 33000 train 0.10110 inner-val 0.10491 best 0.10491@33000 (1371s) ye2026_sepsis=0.1146 imperial_eskd=0.0984 liu2021=0.1017
+2026-10-03T08:19:10 step 33250 train 0.10134 inner-val 0.10491 best 0.10491@33000 (1381s) ye2026_sepsis=0.1146 imperial_eskd=0.0984 liu2021=0.1017
+2026-10-03T08:19:20 step 33500 train 0.10082 inner-val 0.10491 best 0.10491@33000 (1392s) ye2026_sepsis=0.1146 imperial_eskd=0.0984 liu2021=0.1017
+2026-10-03T08:19:30 step 33750 train 0.10121 inner-val 0.10490 best 0.10490@33750 (1402s) ye2026_sepsis=0.1146 imperial_eskd=0.0984 liu2021=0.1017
+2026-10-03T08:19:40 step 34000 train 0.10145 inner-val 0.10490 best 0.10490@34000 (1412s) ye2026_sepsis=0.1146 imperial_eskd=0.0984 liu2021=0.1017
+2026-10-03T08:19:51 step 34250 train 0.10155 inner-val 0.10490 best 0.10490@34000 (1423s) ye2026_sepsis=0.1146 imperial_eskd=0.0984 liu2021=0.1017
+2026-10-03T08:20:01 step 34500 train 0.10157 inner-val 0.10490 best 0.10490@34500 (1433s) ye2026_sepsis=0.1145 imperial_eskd=0.0984 liu2021=0.1017
+2026-10-03T08:20:11 step 34750 train 0.10136 inner-val 0.10489 best 0.10489@34750 (1443s) ye2026_sepsis=0.1145 imperial_eskd=0.0984 liu2021=0.1017
+2026-10-03T08:20:22 step 35000 train 0.10149 inner-val 0.10489 best 0.10489@34750 (1454s) ye2026_sepsis=0.1145 imperial_eskd=0.0984 liu2021=0.1017
+2026-10-03T08:20:32 step 35250 train 0.10144 inner-val 0.10489 best 0.10489@34750 (1464s) ye2026_sepsis=0.1145 imperial_eskd=0.0984 liu2021=0.1017
+2026-10-03T08:20:42 step 35500 train 0.10167 inner-val 0.10489 best 0.10489@34750 (1474s) ye2026_sepsis=0.1145 imperial_eskd=0.0984 liu2021=0.1017
+2026-10-03T08:20:53 step 35750 train 0.10124 inner-val 0.10489 best 0.10489@34750 (1484s) ye2026_sepsis=0.1145 imperial_eskd=0.0984 liu2021=0.1017
+2026-10-03T08:21:03 step 36000 train 0.10153 inner-val 0.10489 best 0.10489@34750 (1495s) ye2026_sepsis=0.1145 imperial_eskd=0.0984 liu2021=0.1017
+2026-10-03T08:21:13 step 36250 train 0.10141 inner-val 0.10488 best 0.10488@36250 (1505s) ye2026_sepsis=0.1145 imperial_eskd=0.0984 liu2021=0.1017
+2026-10-03T08:21:24 step 36500 train 0.10095 inner-val 0.10487 best 0.10487@36500 (1515s) ye2026_sepsis=0.1145 imperial_eskd=0.0984 liu2021=0.1017
+2026-10-03T08:21:34 step 36750 train 0.10156 inner-val 0.10487 best 0.10487@36750 (1526s) ye2026_sepsis=0.1145 imperial_eskd=0.0984 liu2021=0.1017
+2026-10-03T08:21:44 step 37000 train 0.10096 inner-val 0.10486 best 0.10486@37000 (1536s) ye2026_sepsis=0.1145 imperial_eskd=0.0984 liu2021=0.1017
+2026-10-03T08:21:55 step 37250 train 0.10136 inner-val 0.10486 best 0.10486@37000 (1546s) ye2026_sepsis=0.1145 imperial_eskd=0.0984 liu2021=0.1017
+2026-10-03T08:22:05 step 37500 train 0.10139 inner-val 0.10486 best 0.10486@37000 (1557s) ye2026_sepsis=0.1145 imperial_eskd=0.0984 liu2021=0.1017
+2026-10-03T08:22:15 step 37750 train 0.10140 inner-val 0.10486 best 0.10486@37000 (1567s) ye2026_sepsis=0.1145 imperial_eskd=0.0984 liu2021=0.1017
+2026-10-03T08:22:25 step 38000 train 0.10145 inner-val 0.10486 best 0.10486@37000 (1577s) ye2026_sepsis=0.1145 imperial_eskd=0.0984 liu2021=0.1017
+2026-10-03T08:22:37 step 38250 train 0.10149 inner-val 0.10487 best 0.10486@37000 (1588s) ye2026_sepsis=0.1145 imperial_eskd=0.0984 liu2021=0.1017
+2026-10-03T08:22:48 step 38500 train 0.10133 inner-val 0.10486 best 0.10486@37000 (1600s) ye2026_sepsis=0.1145 imperial_eskd=0.0984 liu2021=0.1017
+2026-10-03T08:23:00 step 38750 train 0.10128 inner-val 0.10486 best 0.10486@37000 (1612s) ye2026_sepsis=0.1145 imperial_eskd=0.0984 liu2021=0.1017
+2026-10-03T08:23:12 step 39000 train 0.10118 inner-val 0.10486 best 0.10486@37000 (1624s) ye2026_sepsis=0.1145 imperial_eskd=0.0984 liu2021=0.1017
+2026-10-03T08:23:12 early stop: no improvement in 8 evaluations
+2026-10-03T08:23:16 saved /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_crosscohort_20261002/runs/hoL_immicro/K1000/lora64sat/checkpoint_ft.pt (selected step 37000, inner-val 0.10486 vs pretrained 0.11485); sha 1dc8d1bc8b5f

paper-releases/pbisc-diffusion/updates/20261003T103040Z-crosscohort-k1000-saturation/diffusion/dataset/server_1/model8/plans/paired_crosscohort_20261002/runs/hoL_imperial_eskd/K1000/fullsat/RUN.json

@@ -0,0 +1,546 @@
+{
+ "schema": "pbisc_sdd.ft_crosscohort.run.v1",
+ "source_checkpoint": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/runs/f01_6k_1x_seed20260627/checkpoint_epoch_016.pt",
+ "source_checkpoint_sha256": "8c7cd0e83e2bc19fe03dfcbc084919a1e2f61be8cc8bc12f4caf543dfda91633",
+ "split": "hoL_imperial_eskd",
+ "split_file": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_crosscohort_20261002/splits/hoL_imperial_eskd.json",
+ "split_sha256": "dc497bb9a090f5762cf897e1e1b297cb63e33e459eedfcc3c0a181d4e66b62c4",
+ "held_out": "imperial_eskd",
+ "recipe_k": 1000,
+ "balance": "cohort",
+ "mode": "full",
+ "trainable": {
+  "mode": "full",
+  "n_trainable": 54185184,
+  "n_total_base": 54185184,
+  "trainable_tensors": 64
+ },
+ "lr": 5e-05,
+ "warmup": 200,
+ "batch_size": 256,
+ "ema_decay": 0.999,
+ "max_steps": 200000,
+ "steps_run": 93000,
+ "eval_every": 250,
+ "patience": 8,
+ "selected_step": 91000,
+ "selected_inner_val_loss": 0.10355161361534286,
+ "pretrained_inner_val_loss": 0.11369983355204265,
+ "cohorts": [
+  {
+   "cohort": "immicro",
+   "condition_npz": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_crosscohort_20261002/data/immicro/conditions_raw_K1000.npz",
+   "condition_npz_sha256": "2be5717d19e86bdf38e4838d4bd2886a211ba78bfbfe749c458305a3e0ac2dcc",
+   "train_rows": [
+    "HS1",
+    "HS2",
+    "HS3",
+    "HS6",
+    "HS7",
+    "HS8",
+    "HS9",
+    "HS10",
+    "HS12",
+    "HS13",
+    "HS14",
+    "HS15",
+    "HS16",
+    "HS17",
+    "HS18",
+    "HS19",
+    "HS20",
+    "HS21",
+    "HS23",
+    "HS24",
+    "HS25",
+    "HS28",
+    "HS29",
+    "HS30",
+    "HS32",
+    "HS34",
+    "HS35",
+    "HS36",
+    "HS37",
+    "HS38",
+    "HS39",
+    "HS40",
+    "HS41",
+    "HS42",
+    "HS43",
+    "HS44",
+    "HS45",
+    "HS47",
+    "HS48",
+    "HS49",
+    "HS50",
+    "HS51",
+    "HS52",
+    "HS53",
+    "HS54",
+    "HS56",
+    "HS57",
+    "HS58",
+    "HS59",
+    "HS61",
+    "HS63",
+    "HS64",
+    "HS65",
+    "HS66",
+    "HS67",
+    "HS68",
+    "HS69",
+    "HS70",
+    "HS71",
+    "HS72",
+    "HS73",
+    "HS74",
+    "HS75",
+    "HS76",
+    "HS77",
+    "HS78",
+    "HS79",
+    "HS80",
+    "HS81",
+    "HS82",
+    "HS83",
+    "HS84",
+    "HS85",
+    "HS86",
+    "HS88",
+    "HS89",
+    "HS90",
+    "HS91",
+    "HS92",
+    "HS93",
+    "HS94",
+    "HS95",
+    "HS96",
+    "HS97",
+    "HS98",
+    "HS100",
+    "HS101",
+    "HS102",
+    "HS103",
+    "HS104",
+    "HS105",
+    "HS106",
+    "HS107",
+    "HS110",
+    "HS111"
+   ],
+   "inner_val_rows": [
+    "HS4",
+    "HS11",
+    "HS26",
+    "HS27",
+    "HS31",
+    "HS46",
+    "HS55",
+    "HS60",
+    "HS62",
+    "HS87",
+    "HS108",
+    "HS109"
+   ]
+  },
+  {
+   "cohort": "ye2026_sepsis",
+   "condition_npz": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_crosscohort_20261002/data/ye2026_sepsis/conditions_raw_K1000.npz",
+   "condition_npz_sha256": "ce8b61d7dcd3e62b9e63d52189cfb5bba881f87a661384277a7949a629231cb4",
+   "train_rows": [
+    "Abd-PS04",
+    "Abd-PS05",
+    "Abd-PS07",
+    "Abd-PS19",
+    "Abd-PS20",
+    "Abd-PS27",
+    "Abd-S103",
+    "Abd-S120",
+    "Abd-S123",
+    "Abd-S156",
+    "Abd-S39",
+    "Abd-S42",
+    "Abd-S47",
+    "Abd-S51",
+    "Abd-S57",
+    "Abd-S66",
+    "Abd-S79",
+    "Abd-S83",
+    "HC05",
+    "HC07",
+    "HC09",
+    "HC12",
+    "HC13",
+    "HC16",
+    "IC10",
+    "IC15",
+    "IC16",
+    "IC22",
+    "IC23",
+    "PHC01",
+    "PHC02",
+    "PHC03",
+    "PHC05",
+    "PIC01",
+    "PIC02",
+    "PIC04",
+    "PIC05",
+    "Res-PS02",
+    "Res-PS03",
+    "Res-PS12",
+    "Res-PS21",
+    "Res-PS22",
+    "Res-PS25",
+    "Res-PS26",
+    "Res-S10",
+    "Res-S125",
+    "Res-S127",
+    "Res-S136",
+    "Res-S146",
+    "Res-S149",
+    "Res-S150",
+    "Res-S164",
+    "Res-S166",
+    "Res-S177",
+    "Res-S184",
+    "Res-S19",
+    "Res-S26",
+    "Res-S40",
+    "Res-S77",
+    "Res-S87",
+    "Skin-S140",
+    "Skin-S142",
+    "Skin-S152",
+    "Skin-S38",
+    "Skin-S43",
+    "Skin-S46",
+    "Skin-S52",
+    "Skin-S74",
+    "Uri-S134",
+    "Uri-S147",
+    "Uri-S157",
+    "Uri-S20",
+    "Uri-S93"
+   ],
+   "inner_val_rows": [
+    "Abd-PS23",
+    "Abd-S102",
+    "Abd-S54",
+    "Abd-S98",
+    "HC01",
+    "IC19",
+    "Res-PS16",
+    "Res-S139",
+    "Res-S174"
+   ]
+  },
+  {
+   "cohort": "giroux2022",
+   "condition_npz": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_crosscohort_20261002/data/giroux2022/conditions_raw_K1000.npz",
+   "condition_npz_sha256": "b584655fcfa5fa4c0887d5dd83fcb1c5a0206da4afe5fefab96cba2272b1371c",
+   "train_rows": [
+    "450905_d0",
+    "450905_d14",
+    "0B943B_d0",
+    "0B943B_d14",
+    "7085CA_d0",
+    "7085CA_d7",
+    "7085CA_d14",
+    "2AD75E_d0",
+    "2AD75E_d7",
+    "2AD75E_d14",
+    "B85D75_d0",
+    "B85D75_d7",
+    "180E1A_d0",
+    "180E1A_d7",
+    "180E1A_d14",
+    "82CCF5_d0",
+    "82CCF5_d7",
+    "82CCF5_d14",
+    "BAAF62_d0",
+    "BAAF62_d7",
+    "BAAF62_d14",
+    "0BF51C_d0",
+    "0BF51C_d7",
+    "0BF51C_d14"
+   ],
+   "inner_val_rows": []
+  },
+  {
+   "cohort": "liu2021",
+   "condition_npz": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_crosscohort_20261002/data/liu2021/conditions_raw_K1000.npz",
+   "condition_npz_sha256": "ad3131eabcfb11b166ca37c7585c50abb11efb60e2a736fe62de604794535e0c",
+   "train_rows": [
+    "AA220014_HC",
+    "AA220534_HC",
+    "AA220907_HC",
+    "HGR0000071_T0",
+    "HGR0000074_T0",
+    "HGR0000083_T1",
+    "HGR0000089_T1",
+    "HGR0000091_T0",
+    "HGR0000093_T0",
+    "HGR0000136_T0",
+    "HGR0000137_T0",
+    "HGR0000141_T0",
+    "IM000040_HC",
+    "HGR0000042_T1",
+    "HGR0000069_T0",
+    "HGR0000069_T1",
+    "HGR0000092_T0",
+    "HGR0000094_T0",
+    "HGR0000124_T0",
+    "HGR0000144_T2",
+    "HGR0000391_T0",
+    "HGR0000395_T0",
+    "HGR0000418_T1",
+    "SHD4_HC",
+    "SHD7_HC",
+    "HGR0000051_T0",
+    "HGR0000051_T1",
+    "HGR0000051_T3",
+    "HGR0000101_T1",
+    "HGR0000101_T3",
+    "HGR0000102_T0",
+    "HGR0000102_T1",
+    "HGR0000134_T0",
+    "HGR0000135_T0",
+    "HGR0000135_T1",
+    "HGR0000392_T0",
+    "HGR0000429_T1",
+    "HGR0000429_T2",
+    "SHD3_HC",
+    "SHD5_HC"
+   ],
+   "inner_val_rows": [
+    "HGR0000098_T0",
+    "HGR0000143_T0",
+    "HGR0000115_T1",
+    "SHD1_HC"
+   ]
+  }
+ ],
+ "held_out_test_rows_never_loaded": [
+  "C20_1004",
+  "C20_1404",
+  "C21_1004",
+  "C21_1704",
+  "C21_2504",
+  "C23_2204",
+  "C31_1304",
+  "C31_2004",
+  "C31_2704",
+  "C33_1401",
+  "C33_1601",
+  "C33_1901",
+  "C34_1404",
+  "C34_2104",
+  "C34_2504",
+  "C36_1504",
+  "C36_1704",
+  "C36_2104",
+  "C40_1504",
+  "C40_2204",
+  "C40_0105",
+  "C42_1504",
+  "C42_2204",
+  "C42_2704",
+  "C58_1604",
+  "C58_2501",
+  "C58_2901",
+  "C58_0102",
+  "C58_0402",
+  "C58_1003",
+  "C60_2504",
+  "C63_1704",
+  "C63_2404",
+  "C63_2904",
+  "C65_1704",
+  "C65_2104",
+  "C65_2504",
+  "C69_1804",
+  "C82_2504",
+  "C85_2004",
+  "C93_2104",
+  "C93_2704",
+  "C93_0405",
+  "C101_0601",
+  "C101_1101",
+  "C101_1301",
+  "C101_1801",
+  "C101_2001",
+  "C101_2502",
+  "C103_2204",
+  "C104_2204",
+  "C106_2204",
+  "C108_2204",
+  "C108_1501",
+  "C108_1801",
+  "C108_2201",
+  "C113_0605",
+  "C123_2404",
+  "C123_2904",
+  "C123_0505",
+  "C124_2404",
+  "C124_2704",
+  "C124_0105",
+  "C126_2504",
+  "C126_3004",
+  "C126_0605",
+  "C127_2504",
+  "C127_0405",
+  "C127_1105",
+  "C128_2504",
+  "C128_0405",
+  "C128_1105",
+  "C132_2804",
+  "C132_0205",
+  "C132_1105",
+  "C137_3004",
+  "C137_2502",
+  "C138_3004",
+  "C138_2301",
+  "C138_3001",
+  "C138_0202",
+  "C138_0402",
+  "C140_3004",
+  "C140_2801",
+  "C140_3001",
+  "C140_0202",
+  "C140_0402",
+  "C140_0602",
+  "C140_0902",
+  "C140_1103",
+  "C141_2001",
+  "C141_2201",
+  "C141_3001",
+  "C141_0202",
+  "C142_3004",
+  "C145_3004",
+  "C145_0701",
+  "C145_1201",
+  "C145_1401",
+  "C145_1901",
+  "C145_2301",
+  "C146_3004",
+  "C146_1201",
+  "C146_1401",
+  "C146_1601",
+  "C146_1901",
+  "C146_2301",
+  "C146_2701",
+  "C146_2502",
+  "C147_3004",
+  "C147_1401",
+  "C147_1901",
+  "C152_0405",
+  "C161_1105",
+  "C164_1105",
+  "C167_1205",
+  "C168_1205",
+  "C168_2001",
+  "C168_2301",
+  "C168_3001",
+  "C168_0202",
+  "C168_0403",
+  "C169_1205",
+  "C169_2001",
+  "C169_2301",
+  "C169_0202",
+  "C169_0403",
+  "C170_1205",
+  "C170_2301",
+  "C170_0202",
+  "C170_0402",
+  "C170_0403",
+  "C187_2605",
+  "C187_1201",
+  "C187_1401",
+  "C187_1901",
+  "C187_2301",
+  "C187_2502",
+  "C190_2605",
+  "C190_2901",
+  "C190_0102",
+  "C190_0302",
+  "C190_0502",
+  "C190_0802",
+  "C190_1002",
+  "C190_0903",
+  "C193_2605",
+  "C195_2605",
+  "C197_2605",
+  "C212_2705",
+  "C214_2705",
+  "C234_2805",
+  "C237_2805",
+  "C238_2905",
+  "C240_2905",
+  "C246_2905",
+  "C251_2905"
+ ],
+ "max_abs_weight_change": 1.2548550367355347,
+ "n_tensors_changed": 64,
+ "seed": 20261002,
+ "wall_seconds": 4939.6,
+ "smoke": false,
+ "block": 20000,
+ "saturation_frac": 0.01,
+ "stop_reason": "early_stop",
+ "milestones": [
+  {
+   "milestone_step": 20000,
+   "selected_step": 20000,
+   "selected_inner_val_loss": 0.10430546629208105,
+   "block_improvement": 0.009394367259961595,
+   "total_improvement": 0.009394367259961595,
+   "reason": "block_end",
+   "checkpoint": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_crosscohort_20261002/runs/hoL_imperial_eskd/K1000/fullsat/checkpoint_ft_s20000.pt",
+   "checkpoint_sha256": "07db8e2d15b8d3bf2b3557cd908f0b8ba2cc3a96b0abd00f6b13160f26993ce7"
+  },
+  {
+   "milestone_step": 40000,
+   "selected_step": 40000,
+   "selected_inner_val_loss": 0.10398840397182436,
+   "block_improvement": 0.0003170623202566891,
+   "total_improvement": 0.009711429580218284,
+   "reason": "block_end",
+   "checkpoint": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_crosscohort_20261002/runs/hoL_imperial_eskd/K1000/fullsat/checkpoint_ft_s40000.pt",
+   "checkpoint_sha256": "02862b856805147e73daf92ac391d9753bee0457f9d1d59b2501c3a5f966e384"
+  },
+  {
+   "milestone_step": 60000,
+   "selected_step": 59500,
+   "selected_inner_val_loss": 0.10378229469750766,
+   "block_improvement": 0.00020610927431670423,
+   "total_improvement": 0.009917538854534988,
+   "reason": "block_end",
+   "checkpoint": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_crosscohort_20261002/runs/hoL_imperial_eskd/K1000/fullsat/checkpoint_ft_s60000.pt",
+   "checkpoint_sha256": "69c3d89aa1e760ef42c25bcac20d11767376b9c5cefbf61bd06c183f3645a536"
+  },
+  {
+   "milestone_step": 80000,
+   "selected_step": 79750,
+   "selected_inner_val_loss": 0.10363000631332397,
+   "block_improvement": 0.0001522883841836853,
+   "total_improvement": 0.010069827238718673,
+   "reason": "block_end",
+   "checkpoint": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_crosscohort_20261002/runs/hoL_imperial_eskd/K1000/fullsat/checkpoint_ft_s80000.pt",
+   "checkpoint_sha256": "06c4abbfe2020817b8294999dc77b212f5115779c8ae2c9fcb9cc92b5b13b672"
+  },
+  {
+   "milestone_step": 93000,
+   "selected_step": 91000,
+   "selected_inner_val_loss": 0.10355161361534286,
+   "block_improvement": 7.839269798111392e-05,
+   "total_improvement": 0.010148219936699787,
+   "reason": "early_stop",
+   "checkpoint": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_crosscohort_20261002/runs/hoL_imperial_eskd/K1000/fullsat/checkpoint_ft_s93000.pt",
+   "checkpoint_sha256": "cf699ce234e84fd694e10002de2060fc7097915cadcb392e8a3fa8d40486c1ab"
+  }
+ ],
+ "written_utc": "2026-10-03T07:24:11.532394+00:00",
+ "last_milestone_max_abs_diff": 0.0,
+ "checkpoint_sha256": "704c58eb45cdeacd4d93839c3b3eff991b1f86bab61d21f6972a7d41f9cef59e",
+ "reload_forward_max_abs_diff": 0.0
+}

paper-releases/pbisc-diffusion/updates/20261003T103040Z-crosscohort-k1000-saturation/diffusion/dataset/server_1/model8/plans/paired_crosscohort_20261002/runs/hoL_imperial_eskd/K1000/fullsat/milestones.json

@@ -0,0 +1,52 @@
+[
+ {
+  "milestone_step": 20000,
+  "selected_step": 20000,
+  "selected_inner_val_loss": 0.10430546629208105,
+  "block_improvement": 0.009394367259961595,
+  "total_improvement": 0.009394367259961595,
+  "reason": "block_end",
+  "checkpoint": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_crosscohort_20261002/runs/hoL_imperial_eskd/K1000/fullsat/checkpoint_ft_s20000.pt",
+  "checkpoint_sha256": "07db8e2d15b8d3bf2b3557cd908f0b8ba2cc3a96b0abd00f6b13160f26993ce7"
+ },
+ {
+  "milestone_step": 40000,
+  "selected_step": 40000,
+  "selected_inner_val_loss": 0.10398840397182436,
+  "block_improvement": 0.0003170623202566891,
+  "total_improvement": 0.009711429580218284,
+  "reason": "block_end",
+  "checkpoint": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_crosscohort_20261002/runs/hoL_imperial_eskd/K1000/fullsat/checkpoint_ft_s40000.pt",
+  "checkpoint_sha256": "02862b856805147e73daf92ac391d9753bee0457f9d1d59b2501c3a5f966e384"
+ },
+ {
+  "milestone_step": 60000,
+  "selected_step": 59500,
+  "selected_inner_val_loss": 0.10378229469750766,
+  "block_improvement": 0.00020610927431670423,
+  "total_improvement": 0.009917538854534988,
+  "reason": "block_end",
+  "checkpoint": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_crosscohort_20261002/runs/hoL_imperial_eskd/K1000/fullsat/checkpoint_ft_s60000.pt",
+  "checkpoint_sha256": "69c3d89aa1e760ef42c25bcac20d11767376b9c5cefbf61bd06c183f3645a536"
+ },
+ {
+  "milestone_step": 80000,
+  "selected_step": 79750,
+  "selected_inner_val_loss": 0.10363000631332397,
+  "block_improvement": 0.0001522883841836853,
+  "total_improvement": 0.010069827238718673,
+  "reason": "block_end",
+  "checkpoint": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_crosscohort_20261002/runs/hoL_imperial_eskd/K1000/fullsat/checkpoint_ft_s80000.pt",
+  "checkpoint_sha256": "06c4abbfe2020817b8294999dc77b212f5115779c8ae2c9fcb9cc92b5b13b672"
+ },
+ {
+  "milestone_step": 93000,
+  "selected_step": 91000,
+  "selected_inner_val_loss": 0.10355161361534286,
+  "block_improvement": 7.839269798111392e-05,
+  "total_improvement": 0.010148219936699787,
+  "reason": "early_stop",
+  "checkpoint": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_crosscohort_20261002/runs/hoL_imperial_eskd/K1000/fullsat/checkpoint_ft_s93000.pt",
+  "checkpoint_sha256": "cf699ce234e84fd694e10002de2060fc7097915cadcb392e8a3fa8d40486c1ab"
+ }
+]