Archive 4 checkpoints (1-4)

paper-releases/pbisc-diffusion/updates/20261002T124820Z-crosscohort-liu-default/diffusion/dataset/server_1/model8/plans/paired_crosscohort_20261002/queues/loco_liu_20261002.txt

@@ -0,0 +1,49 @@
+# Cross-cohort (LOCO) benchmark with Liu 2021 in every training pool (user decision 2026-10-02: Liu is the default;
+# splits hoL_*, tools/prepare_loco_extra.py). Replaces queues/loco_20261002.txt, which was stopped on user request
+# after 3 units (ho_giroux2022 / ho_ji2025_chili / ho_ye2026_sepsis, K = 1000, LoRA r16; kept as the no-Liu reference).
+# Priority: K = 1000 LoRA r16, its calibration control, LoRA r64 and full at K = 1000, then K = 2000.
+loco|hoL_giroux2022|1000|lora16|lora|5e-4|--lora-rank|16|--lora-alpha|16
+loco|hoL_ji2025_chili|1000|lora16|lora|5e-4|--lora-rank|16|--lora-alpha|16
+loco|hoL_ye2026_sepsis|1000|lora16|lora|5e-4|--lora-rank|16|--lora-alpha|16
+loco|hoL_immicro|1000|lora16|lora|5e-4|--lora-rank|16|--lora-alpha|16
+loco|hoL_imperial_eskd|1000|lora16|lora|5e-4|--lora-rank|16|--lora-alpha|16
+xcal|hoL_giroux2022|1000
+xcal|hoL_ji2025_chili|1000
+xcal|hoL_ye2026_sepsis|1000
+xcal|hoL_immicro|1000
+xcal|hoL_imperial_eskd|1000
+loco|hoL_giroux2022|1000|lora64|lora|5e-4|--lora-rank|64|--lora-alpha|64
+loco|hoL_ji2025_chili|1000|lora64|lora|5e-4|--lora-rank|64|--lora-alpha|64
+loco|hoL_ye2026_sepsis|1000|lora64|lora|5e-4|--lora-rank|64|--lora-alpha|64
+loco|hoL_immicro|1000|lora64|lora|5e-4|--lora-rank|64|--lora-alpha|64
+loco|hoL_imperial_eskd|1000|lora64|lora|5e-4|--lora-rank|64|--lora-alpha|64
+loco|hoL_giroux2022|1000|full|full|5e-5
+loco|hoL_ji2025_chili|1000|full|full|5e-5
+loco|hoL_ye2026_sepsis|1000|full|full|5e-5
+loco|hoL_immicro|1000|full|full|5e-5
+loco|hoL_imperial_eskd|1000|full|full|5e-5
+zs|giroux2022|2000
+zs|ji2025_chili|2000
+zs|ye2026_sepsis|2000
+zs|immicro|2000
+zs|imperial_eskd|2000
+xcal|hoL_giroux2022|2000
+xcal|hoL_ji2025_chili|2000
+xcal|hoL_ye2026_sepsis|2000
+xcal|hoL_immicro|2000
+xcal|hoL_imperial_eskd|2000
+loco|hoL_giroux2022|2000|lora16|lora|5e-4|--lora-rank|16|--lora-alpha|16
+loco|hoL_ji2025_chili|2000|lora16|lora|5e-4|--lora-rank|16|--lora-alpha|16
+loco|hoL_ye2026_sepsis|2000|lora16|lora|5e-4|--lora-rank|16|--lora-alpha|16
+loco|hoL_immicro|2000|lora16|lora|5e-4|--lora-rank|16|--lora-alpha|16
+loco|hoL_imperial_eskd|2000|lora16|lora|5e-4|--lora-rank|16|--lora-alpha|16
+loco|hoL_giroux2022|2000|lora64|lora|5e-4|--lora-rank|64|--lora-alpha|64
+loco|hoL_ji2025_chili|2000|lora64|lora|5e-4|--lora-rank|64|--lora-alpha|64
+loco|hoL_ye2026_sepsis|2000|lora64|lora|5e-4|--lora-rank|64|--lora-alpha|64
+loco|hoL_immicro|2000|lora64|lora|5e-4|--lora-rank|64|--lora-alpha|64
+loco|hoL_imperial_eskd|2000|lora64|lora|5e-4|--lora-rank|64|--lora-alpha|64
+loco|hoL_giroux2022|2000|full|full|5e-5
+loco|hoL_ji2025_chili|2000|full|full|5e-5
+loco|hoL_ye2026_sepsis|2000|full|full|5e-5
+loco|hoL_immicro|2000|full|full|5e-5
+loco|hoL_imperial_eskd|2000|full|full|5e-5

paper-releases/pbisc-diffusion/updates/20261002T124820Z-crosscohort-liu-default/diffusion/dataset/server_1/model8/plans/paired_crosscohort_20261002/splits/hoL_ji2025_chili/test_xcal_K2000.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d5733ec713c666feb2c66f21cefe9cd1a4a12ff68c75610854b59dd8ebe0c0a8
+size 9833613

paper-releases/pbisc-diffusion/updates/20261002T124820Z-crosscohort-liu-default/diffusion/dataset/server_1/model8/plans/paired_crosscohort_20261002/splits/hoL_ye2026_sepsis/test_xcal_K1000.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b4c2b51f0669e2f0236e7981a6b5b5ddabcd22cbf9e3949384144fb55f5564ed
+size 22639663

paper-releases/pbisc-diffusion/updates/20261002T124820Z-crosscohort-liu-default/diffusion/dataset/server_1/model8/plans/paired_crosscohort_20261002/splits/hoL_ye2026_sepsis/test_xcal_K2000.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:03d9fa67aac63bf8b4a33b220dc4eed8a04b7fc72dbf4c0f6169ebfa24e6a5dd
+size 22639663