Add files using upload-large-folder tool

.gitattributes

@@ -33,3 +33,6 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+experiments/timestep_strategy/figures/error_vs_timestep.png filter=lfs diff=lfs merge=lfs -text
+experiments/timestep_strategy/figures/timestep_strategy_comparison.png filter=lfs diff=lfs merge=lfs -text
+experiments/separation_loss_ablation/figures/separation_loss_ablation_final.png filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,25 @@
+---
+library_name: pytorch
+tags:
+- conditional-diffusion
+- out-of-distribution-detection
+- cifar10
+- thesis
+license: mit
+---
+
+# CIFAR-10 Conditional Diffusion OOD Models
+
+This Hugging Face package contains the model-release side of the CIFAR-10/OOD thesis experiments. Exact final CIFAR checkpoints referenced by the thesis were not present in the local `ssh00` folder; they are therefore documented in `missing_checkpoints.md` rather than replaced by uncertain local weights.
+
+## Reported Thesis Results
+- Within-CIFAR audited AUROC: 98.98%.
+- External OOD audited AUROC range: 90.50% to 96.97% across CIFAR-100, Places365, FashionMNIST, SVHN, and Textures.
+- Separation loss effect: lambda 0.02 reaches 99.03% +/- 0.07% AUROC, versus 80.25% for lambda 0.
+
+## Contents
+- `raw_scores/`: auditable score tensors available locally.
+- `results_json/`: ablation/evaluation JSON artifacts.
+- `tables/` and `figures/`: thesis table and figure artifacts.
+- `dataset_sources.md`: dataset provenance and expected sources.
+- `missing_checkpoints.md`: exact absent checkpoint provenance from thesis documentation.

checksums.sha256

@@ -0,0 +1,74 @@
+94bd424b0dffdd61783b0593d9fdf677abc7fdaaf4f89babab6303f29315c9c8  dataset_sources.md
+24edb189e279970cd95ac5e6e02709eaa068fdec7f4a5fd0d95cb5947237559e  experiments/external_ood/raw_scores/raw_scores_remote_inventory.csv
+eea96e85415d707344682bc9740402ef7536ad20cdfab4f15d18e4ad7169fbc3  experiments/external_ood/raw_scores/seed123_cifar100_scores.pt
+fe6e549a6882414c135e31cb587fe0fa96c1ce8d1499929f26e54fb0f574e2ff  experiments/external_ood/raw_scores/seed123_cifar10_id_scores.pt
+5db6e8d654d19a04903a63a028bf3eee04cbfb2ad11388e38008ed79c779c540  experiments/external_ood/raw_scores/seed123_cifar10_ood_scores.pt
+c36a3aaa549d85e0d051bffd39aaa41f3becd1c5a77d256440cc6cec85eca94a  experiments/external_ood/raw_scores/seed123_fashionmnist_scores.pt
+efea6c9959a2e1eb8c8f45ec805640e2e9788e3f1335adf2c5f8ffd87456406f  experiments/external_ood/raw_scores/seed123_food101_scores.pt
+99f782f861fd1967d3114512f3609567d462c82a3bf94fa50b06e84b9fe8ba37  experiments/external_ood/raw_scores/seed123_stl10_scores.pt
+18467a14e7aa7acb41c6867dac363dc8c01d45aa9f8049e4f7e86f3b0af37c43  experiments/external_ood/raw_scores/seed123_svhn_scores.pt
+3348acbae4bac9d891c40e7f87584ca2003e8185f55a2db31589a11bc7202e6a  experiments/external_ood/raw_scores/seed123_textures_scores.pt
+01da976c1ef147ce098120dc72dca4309b039d3deb0582312f19f450a6b9b923  experiments/external_ood/raw_scores/seed42_cifar100_scores.pt
+f45329e41b35cfb442143379ebdc9d96f488a5f04776ee3d0b4243dec458a757  experiments/external_ood/raw_scores/seed42_cifar10_id_scores.pt
+64599a919c1dc17d4633ad6024d29d9e6f549fbf491420c9dedae234240e1d39  experiments/external_ood/raw_scores/seed42_cifar10_ood_scores.pt
+e2595689d4c135e9968f99117b83739f5805eb520e19378caf6800d7ac1cb037  experiments/external_ood/raw_scores/seed42_fashionmnist_scores.pt
+f7c37c628db9b479196923024de79410d3a5bcf15419698657237df36b2beb1b  experiments/external_ood/raw_scores/seed42_food101_scores.pt
+fa05e2f8e7c80f1dcf3f056aedf5f319e4fb6b87e54ccbd9c6c10326ddab271c  experiments/external_ood/raw_scores/seed42_places365_scores.pt
+63bc986fe2d1750424b3d161056b4deb448580a3665c0265070728173142adad  experiments/external_ood/raw_scores/seed42_stl10_scores.pt
+1d4b841177293ab8ba149b22d031c8af7e79938dcbaf798d4546c396fffba50f  experiments/external_ood/raw_scores/seed42_svhn_scores.pt
+8b5a2cf6d56c7c4ee7ee1db6f0a169ec8af01adaa2d65a0fa24430f3dd9f5d38  experiments/external_ood/raw_scores/seed42_textures_scores.pt
+6f5eff496e89464f01af6dac0a09e9b9ea4c2333cec255654bc9a72c9acc7ba6  experiments/external_ood/raw_scores/seed456_cifar100_scores.pt
+eca890ba2edb578cd2a1f6549295f4212e7287c180e02e8477f751845f7cd579  experiments/external_ood/raw_scores/seed456_cifar10_id_scores.pt
+3d464ff0c432a15cb6c32fce045b7f2088f91ba6a080ff2d4b27a6bb73dea089  experiments/external_ood/raw_scores/seed456_cifar10_ood_scores.pt
+089afd7bc9661058659d07869e2c1400bbce7366df49d4507390d2852fd6692e  experiments/external_ood/raw_scores/seed456_fashionmnist_scores.pt
+5905ea30ce29afc2b3eea511585a3f5e1cc95d1df3a1150fedada9c17f23e545  experiments/external_ood/raw_scores/seed456_food101_scores.pt
+6ebfbb38187ca20426cbdaa3547f81aa5e4d3830a69c6e623d8a06bfe82fbf8f  experiments/external_ood/raw_scores/seed456_stl10_scores.pt
+bbe9b31182ed334606cc8156122ad6f4da2459e90961d7f1c60ae58e271e8107  experiments/external_ood/raw_scores/seed456_svhn_scores.pt
+711b930fb46112f7a5dcfb236947140f0970f65f24524f3bfe80fe25ecdd5adc  experiments/external_ood/raw_scores/seed456_textures_scores.pt
+44136fa355b3678a1146ad16f7e8649e94fb4fc21fe77e8310c060f61caaff8a  experiments/external_ood/results_json/external_ood_results.json
+5ca1db6bdaf482963ef0fe870c0ac658430166ea323f98d5d4396384fb7815b7  experiments/external_ood/tables/external_ood_table.tex
+dd6a6d029a7b6e24d12cfb4c01d0b279f8170f6ac1abf2a93d06adb143be037a  experiments/k_ablation/figures/k_ablation.png
+de1233a80f4884990cb8e26199cb9045ce983d445d1412525fcde04ccc09348d  experiments/k_ablation/results_json/k_ablation_results.json
+452b9391922d95550512e83359e699d6a1cc4a62c0be75220fdbe347794bd7cd  experiments/k_ablation/tables/k_ablation_table.tex
+5b147b2037cd3c01ea0df4a5a207a815eec966ef50cff9b673e61ed56f71b9b6  experiments/main_training/figures/calibration_curves.png
+c5f6fe6eeed8a320d3920d590de92ba42f9069cec328170678b60d21fbd0ba94  experiments/main_training/figures/confusion_matrix.png
+7709d10b29f9bdf00eb996d5e1fc73d628b43dd765cb103d0fc4eb10d98c9465  experiments/main_training/figures/per_class_performance.png
+319619262a6386632c67e6e55c5f49220c73ba3278b17cc067ed5e03d6aefb1f  experiments/main_training/figures/roc_curves_cifar10.png
+93a3552bddfe6dcf7e93e4ee7059bd00b791af7579b6fcf5803327677ebe1f99  experiments/main_training/figures/score_distributions.png
+c38c43e9308fb7206eda29fbaa41cf1a66bfc4f9c8b3072f84eb5877f241b427  experiments/main_training/figures/score_distributions_all.png
+50f96dfd443b95060c4aa444d1846e0b6ce1cc3655558442bd9304396cd88099  experiments/main_training/figures/three_seed_auroc.png
+5f7834a313dc74778c202922da35aec4a65323baf455c40345389e0c179addf1  experiments/main_training/figures/training_curves.png
+cff71f40235d7ea6e6f348b8ed7c40404d72fbae3cd6eb1509920fc2b5459cee  experiments/main_training/tables/main_results_table.tex
+f0a02d1dde9dd5fdc5bf45cdb17ff7fa0772093ea4cd917dbaa251b4ad276268  experiments/scoring_method/figures/scoring_method_comparison.png
+d4a2b5e497467e04e3044fcfd39cf0d4740ca62b65df2a913f4413c1ff902586  experiments/scoring_method/figures/scoring_methods_full.png
+23b5d33e0dda98829b7d451bade41ee79f247c7bedf2730406fb46f0ea9a2c17  experiments/scoring_method/results_json/scoring_method_results.json
+cb2a2e16208af9fad5847b03cd4cfc757ebf897457196bd72798d7e4ef54fdfc  experiments/scoring_method/tables/scoring_method_table.tex
+821da978cc9a33c27cc553dfb604d4acdfcd01b48052d4db781350c1e33ad9b2  experiments/separation_loss_ablation/figures/sep_loss_dual.png
+ac1dd26568bd55e07d36d24851c3b53b5ef25bf3a05c1940d60faee9489a98ff  experiments/separation_loss_ablation/figures/separation_loss_ablation.png
+98bf9835ea7128b6d713cec915473e9b1a2e655e365bc6cf9120ceeae49b9b8e  experiments/separation_loss_ablation/figures/separation_loss_ablation_final.png
+e45576da94da72317429cdfee8f82426ec5f57747def9ecd98840b63d6bc83e8  experiments/separation_loss_ablation/figures/separation_loss_ablation_updated.png
+38f6d1492b523c65546396366080188c44b2b237f93b6d64d72414cb7c7666b2  experiments/separation_loss_ablation/results_json/separation_loss_results.json
+288027d2aee72c4493909db18aed87172acc5858e72c40c85b5cb2e21a9e230b  experiments/separation_loss_ablation/tables/separation_loss_table.tex
+dbb86d9641b596aacfd13cb556e4aafcb96acde80d337d93b7d302fc993fb60c  experiments/timestep_strategy/figures/error_vs_timestep.png
+5931e3b4e6bb84ef99dbb15d8a2a42d37c1a716eeff57c06afe950c3f82578b2  experiments/timestep_strategy/figures/timestep_strategy_comparison.png
+d4ec076286bfd36ba24736dd46ff1eea92b97cd328b70b3e96991d77c92619ae  experiments/timestep_strategy/results_json/timestep_strategy_results.json
+6706e8d4edb6db3a0d5c5e7f4cc51a0192ee619cc02bd96042805940c329be3f  experiments/timestep_strategy/tables/timestep_strategy_table.tex
+e9d220b0b22d564b4ba6de67723082a5828b96c8bf29d8020a1d3236336713bc  missing_checkpoints.md
+791ffac33b2ad634a718320f11a79ddee96ff96b3851ca5233785e053a2f7a57  models/remote_true/main_training/seed123_best_auroc0.9886.ckpt
+305ff58724d0254339a20c7c3ad7e02270b018f2022d7f84afde6ccd73a70ec2  models/remote_true/main_training/seed42_best_auroc0.9873.ckpt
+42ae28da8d69b84440c0362758b0e069954cc278706ad2aec3c892a254889af2  models/remote_true/main_training/seed456_best_auroc0.9887.ckpt
+30200a9a0eec6730bd1b775ad388f8508189278d843343ea53612e33e8606d34  models/remote_true/separation_loss_ablation/sep_loss_0.02_seed456_best_auroc0.9904_ckpt_thesis_auroc0.9903.ckpt
+7e8fc670fa049fe8d74ae0bf4df58c9ab2fc3ea8a22e50e09383136647588539  models/remote_true/separation_loss_ablation/sep_loss_lambda_0p001_2026-02-21_22-03-43_sep_0.001_epoch19_auroc0.9732.ckpt
+5e29012954032dd8c558f67bcf04cfcbd7f2e38398e2737bcbd1c34ea75e16d7  models/remote_true/separation_loss_ablation/sep_loss_lambda_0p01_2026-02-23_18-52-30_sep_0.01_epoch19_auroc0.9869.ckpt
+318a2132b7f934f654ae61d48ee38f71622f750e12118d21c72f4079d99729c6  models/remote_true/separation_loss_ablation/sep_loss_lambda_0p02_2026-02-24_22-57-58_sep_0.02_epoch29_auroc0.9911.ckpt
+64b03548868182d03d44ad675c2648bc39b685a3dbe0a0141f7175c69a6e7239  models/remote_true/separation_loss_ablation/sep_loss_lambda_0p02_seed123_sep_0.02_seed123__2026-02-26_09-43-10_sep_0.02_seed123_epoch29_auroc0.9840.ckpt
+882dbcb4833887158f74db1a650068dee1c8c07c1a584318d6b6080a3f3d9c30  models/remote_true/separation_loss_ablation/sep_loss_lambda_0p02_seed123_sep_0.02_seed123__2026-02-26_22-32-35_sep_0.02_seed123_epoch39_auroc0.9895.ckpt
+6d007ec7a4c0030dcf36c6e6ce828b975c961f87dca05b62f0de83ec148433fc  models/remote_true/separation_loss_ablation/sep_loss_lambda_0p05_2026-02-22_14-33-44_sep_0.05_epoch19_auroc0.9851.ckpt
+7bdee5c15900a7a4db13085afb05dce5dd9540cf6e282f692b2824bd6641421d  models/remote_true/separation_loss_ablation/sep_loss_lambda_0p05_2026-02-23_02-04-16_sep_0.05_epoch19_auroc0.9851.ckpt
+59230ec1bae56c934d6c3fe03e6e07eb7db56af6a11dca2cb4adc78d435b6085  models/remote_true/separation_loss_ablation/sep_loss_lambda_0p0_2026-02-19_06-56-35_sep_0.0_epoch19_auroc0.9781.ckpt
+f8d92995b708316ceacb5c62f2261bb6bbdc6b0ba13ea9b8c4e4a7c7b9c5c2b9  models/remote_true/separation_loss_ablation/sep_loss_lambda_0p0_2026-02-20_16-33-09_sep_0.0_epoch29_auroc0.9648.ckpt
+ca6cb487d94c1d376d4cf98d74cd5a4491ddefbbbe32c47203cd35f91b93eee6  models/remote_true/separation_loss_ablation/sep_loss_lambda_0p0_2026-02-21_05-04-31_sep_0.0_epoch79_auroc0.8025.ckpt
+e483f3be193ac7a96b0dd2a8888891b6cac7a67ad9dbe7cf6ff7abeafae8d49b  models/remote_true/separation_loss_ablation/sep_loss_lambda_0p1_2026-02-22_14-36-51_sep_0.1_epoch149_auroc0.9667.ckpt
+7dcb69acb6fc84cfa54bfe0b501ab42b261d12548fcc064af0eb1888b2c00fef  models/remote_true/separation_loss_ablation/sep_loss_lambda_0p1_2026-02-24_06-00-19_sep_0.1_epoch09_auroc0.8397.ckpt
+2e4d6237fbcad74deede88ba0c4d16ad05a7851e1a806020fd0c10fe25804b83  models/remote_true/separation_loss_ablation/separation_loss_ablation_models.csv
+eb04102a85991698e1f62619853557640040f980e285fc1fc19e5e7f76761563  README.md

dataset_sources.md

@@ -0,0 +1,10 @@
+# Dataset Sources
+
+- CIFAR-10: torchvision dataset, in-distribution class airplane versus non-airplane proxy OOD split.
+- CIFAR-100: torchvision external OOD.
+- SVHN: torchvision SVHN, format 2 cropped.
+- FashionMNIST: torchvision, resized to 32 x 32 and converted to RGB.
+- Places365: external OOD scene dataset.
+- Textures/DTD: Describable Textures Dataset.
+
+The GitHub package contains the evaluation scripts and expected preprocessing logic. Public datasets should be downloaded from their canonical sources during reproduction.

experiments/external_ood/raw_scores/raw_scores_remote_inventory.csv

@@ -0,0 +1,26 @@
+remote_path,local_path,bytes,sha256
+/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/raw_scores/seed123_cifar100_scores.pt,D:\side_hustle\thesis\draft02\reproducibility\cifar10_cdm\huggingface\raw_scores\seed123_cifar100_scores.pt,41324,eea96e85415d707344682bc9740402ef7536ad20cdfab4f15d18e4ad7169fbc3
+/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/raw_scores/seed123_cifar10_id_scores.pt,D:\side_hustle\thesis\draft02\reproducibility\cifar10_cdm\huggingface\raw_scores\seed123_cifar10_id_scores.pt,5302,fe6e549a6882414c135e31cb587fe0fa96c1ce8d1499929f26e54fb0f574e2ff
+/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/raw_scores/seed123_cifar10_ood_scores.pt,D:\side_hustle\thesis\draft02\reproducibility\cifar10_cdm\huggingface\raw_scores\seed123_cifar10_ood_scores.pt,37307,5db6e8d654d19a04903a63a028bf3eee04cbfb2ad11388e38008ed79c779c540
+/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/raw_scores/seed123_fashionmnist_scores.pt,D:\side_hustle\thesis\draft02\reproducibility\cifar10_cdm\huggingface\raw_scores\seed123_fashionmnist_scores.pt,41344,c36a3aaa549d85e0d051bffd39aaa41f3becd1c5a77d256440cc6cec85eca94a
+/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/raw_scores/seed123_food101_scores.pt,D:\side_hustle\thesis\draft02\reproducibility\cifar10_cdm\huggingface\raw_scores\seed123_food101_scores.pt,21287,efea6c9959a2e1eb8c8f45ec805640e2e9788e3f1335adf2c5f8ffd87456406f
+/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/raw_scores/seed123_stl10_scores.pt,D:\side_hustle\thesis\draft02\reproducibility\cifar10_cdm\huggingface\raw_scores\seed123_stl10_scores.pt,21213,99f782f861fd1967d3114512f3609567d462c82a3bf94fa50b06e84b9fe8ba37
+/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/raw_scores/seed123_svhn_scores.pt,D:\side_hustle\thesis\draft02\reproducibility\cifar10_cdm\huggingface\raw_scores\seed123_svhn_scores.pt,105368,18467a14e7aa7acb41c6867dac363dc8c01d45aa9f8049e4f7e86f3b0af37c43
+/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/raw_scores/seed123_textures_scores.pt,D:\side_hustle\thesis\draft02\reproducibility\cifar10_cdm\huggingface\raw_scores\seed123_textures_scores.pt,8812,3348acbae4bac9d891c40e7f87584ca2003e8185f55a2db31589a11bc7202e6a
+/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/raw_scores/seed42_cifar100_scores.pt,D:\side_hustle\thesis\draft02\reproducibility\cifar10_cdm\huggingface\raw_scores\seed42_cifar100_scores.pt,41319,01da976c1ef147ce098120dc72dca4309b039d3deb0582312f19f450a6b9b923
+/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/raw_scores/seed42_cifar10_id_scores.pt,D:\side_hustle\thesis\draft02\reproducibility\cifar10_cdm\huggingface\raw_scores\seed42_cifar10_id_scores.pt,5297,f45329e41b35cfb442143379ebdc9d96f488a5f04776ee3d0b4243dec458a757
+/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/raw_scores/seed42_cifar10_ood_scores.pt,D:\side_hustle\thesis\draft02\reproducibility\cifar10_cdm\huggingface\raw_scores\seed42_cifar10_ood_scores.pt,37302,64599a919c1dc17d4633ad6024d29d9e6f549fbf491420c9dedae234240e1d39
+/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/raw_scores/seed42_fashionmnist_scores.pt,D:\side_hustle\thesis\draft02\reproducibility\cifar10_cdm\huggingface\raw_scores\seed42_fashionmnist_scores.pt,41339,e2595689d4c135e9968f99117b83739f5805eb520e19378caf6800d7ac1cb037
+/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/raw_scores/seed42_food101_scores.pt,D:\side_hustle\thesis\draft02\reproducibility\cifar10_cdm\huggingface\raw_scores\seed42_food101_scores.pt,21218,f7c37c628db9b479196923024de79410d3a5bcf15419698657237df36b2beb1b
+/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/raw_scores/seed42_places365_scores.pt,D:\side_hustle\thesis\draft02\reproducibility\cifar10_cdm\huggingface\raw_scores\seed42_places365_scores.pt,41324,fa05e2f8e7c80f1dcf3f056aedf5f319e4fb6b87e54ccbd9c6c10326ddab271c
+/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/raw_scores/seed42_stl10_scores.pt,D:\side_hustle\thesis\draft02\reproducibility\cifar10_cdm\huggingface\raw_scores\seed42_stl10_scores.pt,21208,63bc986fe2d1750424b3d161056b4deb448580a3665c0265070728173142adad
+/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/raw_scores/seed42_svhn_scores.pt,D:\side_hustle\thesis\draft02\reproducibility\cifar10_cdm\huggingface\raw_scores\seed42_svhn_scores.pt,105363,1d4b841177293ab8ba149b22d031c8af7e79938dcbaf798d4546c396fffba50f
+/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/raw_scores/seed42_textures_scores.pt,D:\side_hustle\thesis\draft02\reproducibility\cifar10_cdm\huggingface\raw_scores\seed42_textures_scores.pt,8807,8b5a2cf6d56c7c4ee7ee1db6f0a169ec8af01adaa2d65a0fa24430f3dd9f5d38
+/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/raw_scores/seed456_cifar100_scores.pt,D:\side_hustle\thesis\draft02\reproducibility\cifar10_cdm\huggingface\raw_scores\seed456_cifar100_scores.pt,41324,6f5eff496e89464f01af6dac0a09e9b9ea4c2333cec255654bc9a72c9acc7ba6
+/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/raw_scores/seed456_cifar10_id_scores.pt,D:\side_hustle\thesis\draft02\reproducibility\cifar10_cdm\huggingface\raw_scores\seed456_cifar10_id_scores.pt,5302,eca890ba2edb578cd2a1f6549295f4212e7287c180e02e8477f751845f7cd579
+/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/raw_scores/seed456_cifar10_ood_scores.pt,D:\side_hustle\thesis\draft02\reproducibility\cifar10_cdm\huggingface\raw_scores\seed456_cifar10_ood_scores.pt,37307,3d464ff0c432a15cb6c32fce045b7f2088f91ba6a080ff2d4b27a6bb73dea089
+/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/raw_scores/seed456_fashionmnist_scores.pt,D:\side_hustle\thesis\draft02\reproducibility\cifar10_cdm\huggingface\raw_scores\seed456_fashionmnist_scores.pt,41344,089afd7bc9661058659d07869e2c1400bbce7366df49d4507390d2852fd6692e
+/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/raw_scores/seed456_food101_scores.pt,D:\side_hustle\thesis\draft02\reproducibility\cifar10_cdm\huggingface\raw_scores\seed456_food101_scores.pt,21287,5905ea30ce29afc2b3eea511585a3f5e1cc95d1df3a1150fedada9c17f23e545
+/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/raw_scores/seed456_stl10_scores.pt,D:\side_hustle\thesis\draft02\reproducibility\cifar10_cdm\huggingface\raw_scores\seed456_stl10_scores.pt,21213,6ebfbb38187ca20426cbdaa3547f81aa5e4d3830a69c6e623d8a06bfe82fbf8f
+/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/raw_scores/seed456_svhn_scores.pt,D:\side_hustle\thesis\draft02\reproducibility\cifar10_cdm\huggingface\raw_scores\seed456_svhn_scores.pt,105368,bbe9b31182ed334606cc8156122ad6f4da2459e90961d7f1c60ae58e271e8107
+/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/raw_scores/seed456_textures_scores.pt,D:\side_hustle\thesis\draft02\reproducibility\cifar10_cdm\huggingface\raw_scores\seed456_textures_scores.pt,8812,711b930fb46112f7a5dcfb236947140f0970f65f24524f3bfe80fe25ecdd5adc

experiments/external_ood/results_json/external_ood_results.json

@@ -0,0 +1 @@
+{}

experiments/external_ood/tables/external_ood_table.tex

@@ -0,0 +1,20 @@
+\begin{table}[htbp]
+\centering
+\caption{External OOD Detection Results ($\lambda=0.01$, K=100, Seed 42). The binary CDM is trained only on CIFAR-10 Airplane (ID) and evaluated against 7 OOD sources of increasing difficulty.}
+\label{tab:external_ood}
+\begin{tabular}{@{}lcc@{}}
+\toprule
+\textbf{OOD Dataset} & \textbf{AUROC} & \textbf{Relative to Within-CIFAR} \\
+\midrule
+Food-101         & 0.9897 & +0.15\% \\
+Within-CIFAR     & 0.9882 & --- \\
+CIFAR-100        & 0.9580 & $-$3.02\% \\
+STL-10           & 0.9426 & $-$4.56\% \\
+FashionMNIST     & 0.9392 & $-$4.90\% \\
+SVHN             & 0.9275 & $-$6.07\% \\
+Textures         & 0.9133 & $-$7.49\% \\
+\midrule
+\textbf{Mean}    & \textbf{0.9512} & --- \\
+\bottomrule
+\end{tabular}
+\end{table}

experiments/k_ablation/results_json/k_ablation_results.json

@@ -0,0 +1,47 @@
+{
+  "within_cifar": {
+    "K_1": {
+      "auroc": 0.9099963333333333,
+      "fpr95": 0.408,
+      "aupr": 0.9877430518385415,
+      "time_seconds": 97.9178569316864,
+      "time_per_sample": 0.00979178569316864
+    },
+    "K_5": {
+      "auroc": 0.9723531666666667,
+      "fpr95": 0.143,
+      "aupr": 0.9965687036233508,
+      "time_seconds": 486.27918767929077,
+      "time_per_sample": 0.04862791876792908
+    },
+    "K_10": {
+      "auroc": 0.9818831111111112,
+      "fpr95": 0.094,
+      "aupr": 0.9977866960333261,
+      "time_seconds": 972.8749701976776,
+      "time_per_sample": 0.09728749701976776
+    },
+    "K_25": {
+      "auroc": 0.9852172222222222,
+      "fpr95": 0.073,
+      "aupr": 0.9982216447717311,
+      "time_seconds": 2431.8085684776306,
+      "time_per_sample": 0.24318085684776305
+    },
+    "K_50": {
+      "auroc": 0.9863586666666667,
+      "fpr95": 0.066,
+      "aupr": 0.998352408323816,
+      "time_seconds": 4861.0760996341705,
+      "time_per_sample": 0.48610760996341706
+    },
+    "K_100": {
+      "auroc": 0.986868,
+      "fpr95": 0.066,
+      "aupr": 0.9983996289418777,
+      "time_seconds": 9723.567284822464,
+      "time_per_sample": 0.9723567284822464
+    }
+  },
+  "svhn": {}
+}

experiments/k_ablation/tables/k_ablation_table.tex

@@ -0,0 +1,21 @@
+% K-Ablation Table (num_trials)
+% Generated by scripts/generate_all_figures_tables.py
+\begin{table}[htbp]
+    \centering
+    \caption{Effect of number of Monte Carlo trials $K$ on OOD scoring.
+             Evaluated on CIFAR-10 binary test set using seed-42 checkpoint.
+             Diminishing returns beyond $K=25$; $K=15$ used during training.}
+    \label{tab:k-ablation}
+    \begin{tabular}{lccc}
+        \toprule
+        $K$ (trials) & AUROC (\%)$\uparrow$ & FPR@95\%TPR$\downarrow$ & Time (s) \\
+        \midrule
+        1 & 91.00 & 40.8 & 97.9 \\
+        5 & 97.24 & 14.3 & 486.3 \\
+        10 & 98.19 & 9.4 & 972.9 \\
+        25 & \textbf{98.52} & \textbf{7.3} & 2431.8 \\
+        50 & \textbf{98.64} & \textbf{6.6} & 4861.1 \\
+        100 & \textbf{98.69} & \textbf{6.6} & 9723.6 \\
+        \bottomrule
+    \end{tabular}
+\end{table}

experiments/main_training/tables/main_results_table.tex

@@ -0,0 +1,24 @@
+% Main Results Table — CDM Binary OOD Detection
+% Generated by scripts/generate_all_figures_tables.py
+\begin{table}[htbp]
+    \centering
+    \caption{Binary CDM OOD detection performance (AUROC, FPR@95\%TPR).
+             Within-dataset result is mean $\pm$ std over 3 seeds (42/123/456).
+             External OOD is mean $\pm$ std evaluated on the seed-42 checkpoint
+             (K=100 MC trials, difference scoring).}
+    \label{tab:main-results}
+    \begin{tabular}{lccc}
+        \toprule
+        OOD Dataset & AUROC (\%)$\uparrow$ & FPR95 (\%)$\downarrow$ & AUPR$\uparrow$ \\
+        \midrule
+        CIFAR-10 (Airplane vs Others) & 98.82 $\pm$ 0.06 & — & — \\
+        SVHN & — & — & — \\
+        CIFAR-100 & — & — & — \\
+        Textures (DTD) & — & — & — \\
+        FashionMNIST & — & — & — \\
+        STL-10 & — & — & — \\
+        Food-101 & — & — & — \\
+        Places365 & — & — & — \\
+        \bottomrule
+    \end{tabular}
+\end{table}

experiments/scoring_method/results_json/scoring_method_results.json

@@ -0,0 +1,36 @@
+{
+  "within_cifar": {
+    "difference": {
+      "auroc": 0.9869498333333332,
+      "fpr95": 0.063,
+      "aupr": 0.9984283144808994
+    },
+    "ratio": {
+      "auroc": 0.9862080555555556,
+      "fpr95": 0.066,
+      "aupr": 0.9982978617155271
+    },
+    "id_error": {
+      "auroc": 0.7830467777777776,
+      "fpr95": 0.67,
+      "aupr": 0.9653815554396192
+    }
+  },
+  "svhn": {
+    "difference": {
+      "auroc": 0.9413390058389677,
+      "fpr95": 0.196,
+      "aupr": 0.9966887629225921
+    },
+    "ratio": {
+      "auroc": 0.9606347572218807,
+      "fpr95": 0.225,
+      "aupr": 0.9983185536493355
+    },
+    "id_error": {
+      "auroc": 0.2023230831284573,
+      "fpr95": 0.965,
+      "aupr": 0.908071538261028
+    }
+  }
+}

experiments/scoring_method/tables/scoring_method_table.tex

@@ -0,0 +1,20 @@
+% Scoring Method Comparison Table
+% Generated by scripts/generate_all_figures_tables.py
+\begin{table}[htbp]
+    \centering
+    \caption{Comparison of OOD scoring methods (seed-42 checkpoint, K=50 trials).
+             \texttt{difference} and \texttt{ratio} perform similarly within CIFAR-10;
+             \texttt{ratio} marginally better on external SVHN OOD.
+             \texttt{id\_error} (ID-only scoring without class conditioning) is much worse,
+             confirming binary conditioning is essential.}
+    \label{tab:scoring-methods}
+    \begin{tabular}{lccc}
+        \toprule
+        Scoring Method & CIFAR AUROC (\%)$\uparrow$ & CIFAR FPR95 (\%)$\downarrow$ & SVHN AUROC (\%)$\uparrow$ \\
+        \midrule
+        \texttt{difference} & \textbf{98.69} & 6.3 & 94.1 \\
+        \texttt{ratio} & 98.62 & 6.6 & \textbf{96.1} \\
+        \texttt{id\_error} & 78.30 & 67.0 & 20.2 \\
+        \bottomrule
+    \end{tabular}
+\end{table}

experiments/separation_loss_ablation/results_json/separation_loss_results.json

@@ -0,0 +1,21 @@
+{
+  "weights": [
+    0.0,
+    0.001,
+    0.01,
+    0.05,
+    0.1
+  ],
+  "within_cifar": {
+    "0.01": {
+      "auroc": 0.9865895000000001,
+      "fpr95": 0.064
+    }
+  },
+  "svhn": {
+    "0.01": {
+      "auroc": 0.8943492240319606,
+      "fpr95": 0.285
+    }
+  }
+}

experiments/separation_loss_ablation/tables/separation_loss_table.tex

@@ -0,0 +1,24 @@
+% Separation Loss Ablation Table
+% Generated by scripts/generate_all_figures_tables.py
+% Note: SVHN cols only populated for lambda=0.01 (from scoring_method_results.json).
+%       Full external OOD evaluation pending completion.
+\begin{table}[htbp]
+    \centering
+    \caption{Effect of separation loss weight $\lambda$ on CIFAR-10 OOD detection
+             (seed=42, K=15 MC trials during training, K=100 for CIFAR final AUROC).
+             Best epoch reports early stopping result.
+             SVHN column only available for $\lambda=0.01$ (see ablation script).}
+    \label{tab:separation-loss}
+    \begin{tabular}{lcccc}
+        \toprule
+        $\lambda$ & CIFAR AUROC (\%) & Best Epoch & SVHN AUROC (\%) & SVHN FPR95 (\%) \\
+        \midrule
+        0.0 & 80.25 & 79 & — & — \\
+        0.001 & 97.32 & 19 & — & — \\
+        0.01 & 98.69 & 19 & 94.1 & 28.5 \\
+        0.02 & \textbf{99.11} & 29 & — & — \\ % \textbf{Best AUROC}
+        0.05 & 98.51 & 19 & — & — \\
+        0.1 & 96.67 & 149 & — & — \\
+        \bottomrule
+    \end{tabular}
+\end{table}

experiments/timestep_strategy/results_json/timestep_strategy_results.json

@@ -0,0 +1,367 @@
+{
+  "within_cifar": {
+    "uniform": {
+      "auroc": 0.9886852222222222,
+      "fpr95": 0.054,
+      "aupr": 0.9986370983243654
+    },
+    "mid_focus": {
+      "auroc": 0.9854892222222222,
+      "fpr95": 0.076,
+      "aupr": 0.9982258725401427
+    },
+    "stratified": {
+      "auroc": 0.9881399444444445,
+      "fpr95": 0.059,
+      "aupr": 0.9985863517967093
+    }
+  },
+  "svhn": {
+    "uniform": {
+      "auroc": 0.9543605562384757,
+      "fpr95": 0.146,
+      "aupr": 0.9974213022601522
+    },
+    "mid_focus": {
+      "auroc": 0.9380184004302397,
+      "fpr95": 0.197,
+      "aupr": 0.996371581168088
+    },
+    "stratified": {
+      "auroc": 0.9498002650583898,
+      "fpr95": 0.163,
+      "aupr": 0.9972195075319602
+    }
+  },
+  "per_timestep": {
+    "timesteps": [
+      50,
+      100,
+      150,
+      200,
+      250,
+      300,
+      350,
+      400,
+      450,
+      500,
+      550,
+      600,
+      650,
+      700,
+      750,
+      800,
+      850,
+      900,
+      950
+    ],
+    "id": {
+      "c0_t50": {
+        "mean": 0.16123446822166443,
+        "std": 0.05786355584859848
+      },
+      "c0_t100": {
+        "mean": 0.10442730039358139,
+        "std": 0.03873381391167641
+      },
+      "c0_t150": {
+        "mean": 0.0775127112865448,
+        "std": 0.02978641912341118
+      },
+      "c0_t200": {
+        "mean": 0.06270774453878403,
+        "std": 0.024262018501758575
+      },
+      "c0_t250": {
+        "mean": 0.05240098759531975,
+        "std": 0.02019326575100422
+      },
+      "c0_t300": {
+        "mean": 0.04459163546562195,
+        "std": 0.01662287674844265
+      },
+      "c0_t350": {
+        "mean": 0.03879544511437416,
+        "std": 0.014503664337098598
+      },
+      "c0_t400": {
+        "mean": 0.033743977546691895,
+        "std": 0.01320528332144022
+      },
+      "c0_t450": {
+        "mean": 0.029082385823130608,
+        "std": 0.010959829203784466
+      },
+      "c0_t500": {
+        "mean": 0.026250595226883888,
+        "std": 0.009717644192278385
+      },
+      "c0_t550": {
+        "mean": 0.022424820810556412,
+        "std": 0.008256726898252964
+      },
+      "c0_t600": {
+        "mean": 0.019210368394851685,
+        "std": 0.006838190834969282
+      },
+      "c0_t650": {
+        "mean": 0.01615460403263569,
+        "std": 0.006200442556291819
+      },
+      "c0_t700": {
+        "mean": 0.013310677371919155,
+        "std": 0.005098859779536724
+      },
+      "c0_t750": {
+        "mean": 0.010465993545949459,
+        "std": 0.0037172066513448954
+      },
+      "c0_t800": {
+        "mean": 0.008377683348953724,
+        "std": 0.0031924911309033632
+      },
+      "c0_t850": {
+        "mean": 0.005653513129800558,
+        "std": 0.002224632538855076
+      },
+      "c0_t900": {
+        "mean": 0.003275742521509528,
+        "std": 0.0013118572533130646
+      },
+      "c0_t950": {
+        "mean": 0.0013495800085365772,
+        "std": 0.0005494702490977943
+      },
+      "c1_t50": {
+        "mean": 0.1606561243534088,
+        "std": 0.05567626282572746
+      },
+      "c1_t100": {
+        "mean": 0.10685943067073822,
+        "std": 0.03887505829334259
+      },
+      "c1_t150": {
+        "mean": 0.07943448424339294,
+        "std": 0.029171813279390335
+      },
+      "c1_t200": {
+        "mean": 0.06453787535429001,
+        "std": 0.023446492850780487
+      },
+      "c1_t250": {
+        "mean": 0.05437518656253815,
+        "std": 0.0198157150298357
+      },
+      "c1_t300": {
+        "mean": 0.04662742465734482,
+        "std": 0.01747937500476837
+      },
+      "c1_t350": {
+        "mean": 0.040038056671619415,
+        "std": 0.014310726895928383
+      },
+      "c1_t400": {
+        "mean": 0.034622400999069214,
+        "std": 0.012737436220049858
+      },
+      "c1_t450": {
+        "mean": 0.030930055305361748,
+        "std": 0.012175540439784527
+      },
+      "c1_t500": {
+        "mean": 0.02622835338115692,
+        "std": 0.00925239734351635
+      },
+      "c1_t550": {
+        "mean": 0.02286655828356743,
+        "std": 0.008599987253546715
+      },
+      "c1_t600": {
+        "mean": 0.019960874691605568,
+        "std": 0.007759998086839914
+      },
+      "c1_t650": {
+        "mean": 0.017022253945469856,
+        "std": 0.006253428291529417
+      },
+      "c1_t700": {
+        "mean": 0.013848936185240746,
+        "std": 0.005125450901687145
+      },
+      "c1_t750": {
+        "mean": 0.010915481485426426,
+        "std": 0.003997159190475941
+      },
+      "c1_t800": {
+        "mean": 0.008292279206216335,
+        "std": 0.0029704829212278128
+      },
+      "c1_t850": {
+        "mean": 0.0059583126567304134,
+        "std": 0.002324610250070691
+      },
+      "c1_t900": {
+        "mean": 0.0035888850688934326,
+        "std": 0.0013977768830955029
+      },
+      "c1_t950": {
+        "mean": 0.001462708692997694,
+        "std": 0.00044951261952519417
+      }
+    },
+    "ood": {
+      "c0_t50": {
+        "mean": 0.22906392812728882,
+        "std": 0.06499648094177246
+      },
+      "c0_t100": {
+        "mean": 0.15222781896591187,
+        "std": 0.04274160414934158
+      },
+      "c0_t150": {
+        "mean": 0.11631786823272705,
+        "std": 0.03303564339876175
+      },
+      "c0_t200": {
+        "mean": 0.09286506474018097,
+        "std": 0.026360269635915756
+      },
+      "c0_t250": {
+        "mean": 0.07722136378288269,
+        "std": 0.022630464285612106
+      },
+      "c0_t300": {
+        "mean": 0.06502427160739899,
+        "std": 0.0195158738642931
+      },
+      "c0_t350": {
+        "mean": 0.05567947402596474,
+        "std": 0.016512945294380188
+      },
+      "c0_t400": {
+        "mean": 0.04731416702270508,
+        "std": 0.013769409619271755
+      },
+      "c0_t450": {
+        "mean": 0.04148917272686958,
+        "std": 0.012854596599936485
+      },
+      "c0_t500": {
+        "mean": 0.034694258123636246,
+        "std": 0.010574056766927242
+      },
+      "c0_t550": {
+        "mean": 0.03002982586622238,
+        "std": 0.009448406286537647
+      },
+      "c0_t600": {
+        "mean": 0.025195691734552383,
+        "std": 0.007569901645183563
+      },
+      "c0_t650": {
+        "mean": 0.020711131393909454,
+        "std": 0.007039411459118128
+      },
+      "c0_t700": {
+        "mean": 0.016733329743146896,
+        "std": 0.00529054319486022
+      },
+      "c0_t750": {
+        "mean": 0.013030176982283592,
+        "std": 0.004108891822397709
+      },
+      "c0_t800": {
+        "mean": 0.009448970668017864,
+        "std": 0.0030608815141022205
+      },
+      "c0_t850": {
+        "mean": 0.00628605717793107,
+        "std": 0.002177068032324314
+      },
+      "c0_t900": {
+        "mean": 0.003598423209041357,
+        "std": 0.0012530161766335368
+      },
+      "c0_t950": {
+        "mean": 0.0013986843405291438,
+        "std": 0.0004934321623295546
+      },
+      "c1_t50": {
+        "mean": 0.22520293295383453,
+        "std": 0.06432735174894333
+      },
+      "c1_t100": {
+        "mean": 0.1499990075826645,
+        "std": 0.04188496619462967
+      },
+      "c1_t150": {
+        "mean": 0.11321725696325302,
+        "std": 0.03249520808458328
+      },
+      "c1_t200": {
+        "mean": 0.08936502039432526,
+        "std": 0.025120839476585388
+      },
+      "c1_t250": {
+        "mean": 0.07322543114423752,
+        "std": 0.021838413551449776
+      },
+      "c1_t300": {
+        "mean": 0.06204748898744583,
+        "std": 0.018268465995788574
+      },
+      "c1_t350": {
+        "mean": 0.05241883546113968,
+        "std": 0.01606724224984646
+      },
+      "c1_t400": {
+        "mean": 0.0449826605618,
+        "std": 0.01481825951486826
+      },
+      "c1_t450": {
+        "mean": 0.03945695608854294,
+        "std": 0.012820725329220295
+      },
+      "c1_t500": {
+        "mean": 0.0332314670085907,
+        "std": 0.010059903375804424
+      },
+      "c1_t550": {
+        "mean": 0.02762959897518158,
+        "std": 0.009240682236850262
+      },
+      "c1_t600": {
+        "mean": 0.023351749405264854,
+        "std": 0.007416155654937029
+      },
+      "c1_t650": {
+        "mean": 0.0198295246809721,
+        "std": 0.006654239259660244
+      },
+      "c1_t700": {
+        "mean": 0.015598704107105732,
+        "std": 0.00491707818582654
+      },
+      "c1_t750": {
+        "mean": 0.01249911542981863,
+        "std": 0.004252676852047443
+      },
+      "c1_t800": {
+        "mean": 0.009655343368649483,
+        "std": 0.0032796640880405903
+      },
+      "c1_t850": {
+        "mean": 0.00628657266497612,
+        "std": 0.0020485452841967344
+      },
+      "c1_t900": {
+        "mean": 0.003460217034444213,
+        "std": 0.0012720918748527765
+      },
+      "c1_t950": {
+        "mean": 0.001465486828237772,
+        "std": 0.0005407508579082787
+      }
+    }
+  }
+}

experiments/timestep_strategy/tables/timestep_strategy_table.tex

@@ -0,0 +1,15 @@
+% Timestep Strategy Table
+\begin{table}[htbp]
+    \centering
+    \caption{Comparison of timestep sampling strategies.}
+    \label{tab:timestep-strategy}
+    \begin{tabular}{lccc}
+        \toprule
+        Strategy & CIFAR AUROC (\%) & SVHN AUROC (\%) \\
+        \midrule
+        Uniform & 98.9 & 95.4 \\
+        Mid Focus & 98.5 & 93.8 \\
+        Stratified & 98.8 & 95.0 \\
+        \bottomrule
+    \end{tabular}
+\end{table}

manifest.csv

@@ -0,0 +1,75 @@
+path,bytes,sha256
+dataset_sources.md,543,94bd424b0dffdd61783b0593d9fdf677abc7fdaaf4f89babab6303f29315c9c8
+experiments/external_ood/raw_scores/raw_scores_remote_inventory.csv,7274,24edb189e279970cd95ac5e6e02709eaa068fdec7f4a5fd0d95cb5947237559e
+experiments/external_ood/raw_scores/seed123_cifar100_scores.pt,41324,eea96e85415d707344682bc9740402ef7536ad20cdfab4f15d18e4ad7169fbc3
+experiments/external_ood/raw_scores/seed123_cifar10_id_scores.pt,5302,fe6e549a6882414c135e31cb587fe0fa96c1ce8d1499929f26e54fb0f574e2ff
+experiments/external_ood/raw_scores/seed123_cifar10_ood_scores.pt,37307,5db6e8d654d19a04903a63a028bf3eee04cbfb2ad11388e38008ed79c779c540
+experiments/external_ood/raw_scores/seed123_fashionmnist_scores.pt,41344,c36a3aaa549d85e0d051bffd39aaa41f3becd1c5a77d256440cc6cec85eca94a
+experiments/external_ood/raw_scores/seed123_food101_scores.pt,21287,efea6c9959a2e1eb8c8f45ec805640e2e9788e3f1335adf2c5f8ffd87456406f
+experiments/external_ood/raw_scores/seed123_stl10_scores.pt,21213,99f782f861fd1967d3114512f3609567d462c82a3bf94fa50b06e84b9fe8ba37
+experiments/external_ood/raw_scores/seed123_svhn_scores.pt,105368,18467a14e7aa7acb41c6867dac363dc8c01d45aa9f8049e4f7e86f3b0af37c43
+experiments/external_ood/raw_scores/seed123_textures_scores.pt,8812,3348acbae4bac9d891c40e7f87584ca2003e8185f55a2db31589a11bc7202e6a
+experiments/external_ood/raw_scores/seed42_cifar100_scores.pt,41319,01da976c1ef147ce098120dc72dca4309b039d3deb0582312f19f450a6b9b923
+experiments/external_ood/raw_scores/seed42_cifar10_id_scores.pt,5297,f45329e41b35cfb442143379ebdc9d96f488a5f04776ee3d0b4243dec458a757
+experiments/external_ood/raw_scores/seed42_cifar10_ood_scores.pt,37302,64599a919c1dc17d4633ad6024d29d9e6f549fbf491420c9dedae234240e1d39
+experiments/external_ood/raw_scores/seed42_fashionmnist_scores.pt,41339,e2595689d4c135e9968f99117b83739f5805eb520e19378caf6800d7ac1cb037
+experiments/external_ood/raw_scores/seed42_food101_scores.pt,21218,f7c37c628db9b479196923024de79410d3a5bcf15419698657237df36b2beb1b
+experiments/external_ood/raw_scores/seed42_places365_scores.pt,41324,fa05e2f8e7c80f1dcf3f056aedf5f319e4fb6b87e54ccbd9c6c10326ddab271c
+experiments/external_ood/raw_scores/seed42_stl10_scores.pt,21208,63bc986fe2d1750424b3d161056b4deb448580a3665c0265070728173142adad
+experiments/external_ood/raw_scores/seed42_svhn_scores.pt,105363,1d4b841177293ab8ba149b22d031c8af7e79938dcbaf798d4546c396fffba50f
+experiments/external_ood/raw_scores/seed42_textures_scores.pt,8807,8b5a2cf6d56c7c4ee7ee1db6f0a169ec8af01adaa2d65a0fa24430f3dd9f5d38
+experiments/external_ood/raw_scores/seed456_cifar100_scores.pt,41324,6f5eff496e89464f01af6dac0a09e9b9ea4c2333cec255654bc9a72c9acc7ba6
+experiments/external_ood/raw_scores/seed456_cifar10_id_scores.pt,5302,eca890ba2edb578cd2a1f6549295f4212e7287c180e02e8477f751845f7cd579
+experiments/external_ood/raw_scores/seed456_cifar10_ood_scores.pt,37307,3d464ff0c432a15cb6c32fce045b7f2088f91ba6a080ff2d4b27a6bb73dea089
+experiments/external_ood/raw_scores/seed456_fashionmnist_scores.pt,41344,089afd7bc9661058659d07869e2c1400bbce7366df49d4507390d2852fd6692e
+experiments/external_ood/raw_scores/seed456_food101_scores.pt,21287,5905ea30ce29afc2b3eea511585a3f5e1cc95d1df3a1150fedada9c17f23e545
+experiments/external_ood/raw_scores/seed456_stl10_scores.pt,21213,6ebfbb38187ca20426cbdaa3547f81aa5e4d3830a69c6e623d8a06bfe82fbf8f
+experiments/external_ood/raw_scores/seed456_svhn_scores.pt,105368,bbe9b31182ed334606cc8156122ad6f4da2459e90961d7f1c60ae58e271e8107
+experiments/external_ood/raw_scores/seed456_textures_scores.pt,8812,711b930fb46112f7a5dcfb236947140f0970f65f24524f3bfe80fe25ecdd5adc
+experiments/external_ood/results_json/external_ood_results.json,2,44136fa355b3678a1146ad16f7e8649e94fb4fc21fe77e8310c060f61caaff8a
+experiments/external_ood/tables/external_ood_table.tex,742,5ca1db6bdaf482963ef0fe870c0ac658430166ea323f98d5d4396384fb7815b7
+experiments/k_ablation/figures/k_ablation.png,126167,dd6a6d029a7b6e24d12cfb4c01d0b279f8170f6ac1abf2a93d06adb143be037a
+experiments/k_ablation/results_json/k_ablation_results.json,1217,de1233a80f4884990cb8e26199cb9045ce983d445d1412525fcde04ccc09348d
+experiments/k_ablation/tables/k_ablation_table.tex,842,452b9391922d95550512e83359e699d6a1cc4a62c0be75220fdbe347794bd7cd
+experiments/main_training/figures/calibration_curves.png,141264,5b147b2037cd3c01ea0df4a5a207a815eec966ef50cff9b673e61ed56f71b9b6
+experiments/main_training/figures/confusion_matrix.png,78998,c5f6fe6eeed8a320d3920d590de92ba42f9069cec328170678b60d21fbd0ba94
+experiments/main_training/figures/per_class_performance.png,162036,7709d10b29f9bdf00eb996d5e1fc73d628b43dd765cb103d0fc4eb10d98c9465
+experiments/main_training/figures/roc_curves_cifar10.png,181333,319619262a6386632c67e6e55c5f49220c73ba3278b17cc067ed5e03d6aefb1f
+experiments/main_training/figures/score_distributions.png,229117,93a3552bddfe6dcf7e93e4ee7059bd00b791af7579b6fcf5803327677ebe1f99
+experiments/main_training/figures/score_distributions_all.png,468037,c38c43e9308fb7206eda29fbaa41cf1a66bfc4f9c8b3072f84eb5877f241b427
+experiments/main_training/figures/three_seed_auroc.png,64055,50f96dfd443b95060c4aa444d1846e0b6ce1cc3655558442bd9304396cd88099
+experiments/main_training/figures/training_curves.png,214652,5f7834a313dc74778c202922da35aec4a65323baf455c40345389e0c179addf1
+experiments/main_training/tables/main_results_table.tex,995,cff71f40235d7ea6e6f348b8ed7c40404d72fbae3cd6eb1509920fc2b5459cee
+experiments/scoring_method/figures/scoring_method_comparison.png,90004,f0a02d1dde9dd5fdc5bf45cdb17ff7fa0772093ea4cd917dbaa251b4ad276268
+experiments/scoring_method/figures/scoring_methods_full.png,85703,d4a2b5e497467e04e3044fcfd39cf0d4740ca62b65df2a913f4413c1ff902586
+experiments/scoring_method/results_json/scoring_method_results.json,728,23b5d33e0dda98829b7d451bade41ee79f247c7bedf2730406fb46f0ea9a2c17
+experiments/scoring_method/tables/scoring_method_table.tex,936,cb2a2e16208af9fad5847b03cd4cfc757ebf897457196bd72798d7e4ef54fdfc
+experiments/separation_loss_ablation/figures/sep_loss_dual.png,115468,821da978cc9a33c27cc553dfb604d4acdfcd01b48052d4db781350c1e33ad9b2
+experiments/separation_loss_ablation/figures/separation_loss_ablation.png,90207,ac1dd26568bd55e07d36d24851c3b53b5ef25bf3a05c1940d60faee9489a98ff
+experiments/separation_loss_ablation/figures/separation_loss_ablation_final.png,107127,98bf9835ea7128b6d713cec915473e9b1a2e655e365bc6cf9120ceeae49b9b8e
+experiments/separation_loss_ablation/figures/separation_loss_ablation_updated.png,100459,e45576da94da72317429cdfee8f82426ec5f57747def9ecd98840b63d6bc83e8
+experiments/separation_loss_ablation/results_json/separation_loss_results.json,264,38f6d1492b523c65546396366080188c44b2b237f93b6d64d72414cb7c7666b2
+experiments/separation_loss_ablation/tables/separation_loss_table.tex,1080,288027d2aee72c4493909db18aed87172acc5858e72c40c85b5cb2e21a9e230b
+experiments/timestep_strategy/figures/error_vs_timestep.png,284649,dbb86d9641b596aacfd13cb556e4aafcb96acde80d337d93b7d302fc993fb60c
+experiments/timestep_strategy/figures/timestep_strategy_comparison.png,103511,5931e3b4e6bb84ef99dbb15d8a2a42d37c1a716eeff57c06afe950c3f82578b2
+experiments/timestep_strategy/results_json/timestep_strategy_results.json,8722,d4ec076286bfd36ba24736dd46ff1eea92b97cd328b70b3e96991d77c92619ae
+experiments/timestep_strategy/tables/timestep_strategy_table.tex,423,6706e8d4edb6db3a0d5c5e7f4cc51a0192ee619cc02bd96042805940c329be3f
+missing_checkpoints.md,1742,e9d220b0b22d564b4ba6de67723082a5828b96c8bf29d8020a1d3236336713bc
+models/remote_true/main_training/seed123_best_auroc0.9886.ckpt,825944028,791ffac33b2ad634a718320f11a79ddee96ff96b3851ca5233785e053a2f7a57
+models/remote_true/main_training/seed42_best_auroc0.9873.ckpt,825944028,305ff58724d0254339a20c7c3ad7e02270b018f2022d7f84afde6ccd73a70ec2
+models/remote_true/main_training/seed456_best_auroc0.9887.ckpt,825944028,42ae28da8d69b84440c0362758b0e069954cc278706ad2aec3c892a254889af2
+models/remote_true/separation_loss_ablation/sep_loss_0.02_seed456_best_auroc0.9904_ckpt_thesis_auroc0.9903.ckpt,825944156,30200a9a0eec6730bd1b775ad388f8508189278d843343ea53612e33e8606d34
+models/remote_true/separation_loss_ablation/sep_loss_lambda_0p001_2026-02-21_22-03-43_sep_0.001_epoch19_auroc0.9732.ckpt,825944092,7e8fc670fa049fe8d74ae0bf4df58c9ab2fc3ea8a22e50e09383136647588539
+models/remote_true/separation_loss_ablation/sep_loss_lambda_0p01_2026-02-23_18-52-30_sep_0.01_epoch19_auroc0.9869.ckpt,825944028,5e29012954032dd8c558f67bcf04cfcbd7f2e38398e2737bcbd1c34ea75e16d7
+models/remote_true/separation_loss_ablation/sep_loss_lambda_0p02_2026-02-24_22-57-58_sep_0.02_epoch29_auroc0.9911.ckpt,825944028,318a2132b7f934f654ae61d48ee38f71622f750e12118d21c72f4079d99729c6
+models/remote_true/separation_loss_ablation/sep_loss_lambda_0p02_seed123_sep_0.02_seed123__2026-02-26_09-43-10_sep_0.02_seed123_epoch29_auroc0.9840.ckpt,825944156,64b03548868182d03d44ad675c2648bc39b685a3dbe0a0141f7175c69a6e7239
+models/remote_true/separation_loss_ablation/sep_loss_lambda_0p02_seed123_sep_0.02_seed123__2026-02-26_22-32-35_sep_0.02_seed123_epoch39_auroc0.9895.ckpt,825944156,882dbcb4833887158f74db1a650068dee1c8c07c1a584318d6b6080a3f3d9c30
+models/remote_true/separation_loss_ablation/sep_loss_lambda_0p05_2026-02-22_14-33-44_sep_0.05_epoch19_auroc0.9851.ckpt,825944028,6d007ec7a4c0030dcf36c6e6ce828b975c961f87dca05b62f0de83ec148433fc
+models/remote_true/separation_loss_ablation/sep_loss_lambda_0p05_2026-02-23_02-04-16_sep_0.05_epoch19_auroc0.9851.ckpt,825944028,7bdee5c15900a7a4db13085afb05dce5dd9540cf6e282f692b2824bd6641421d
+models/remote_true/separation_loss_ablation/sep_loss_lambda_0p0_2026-02-19_06-56-35_sep_0.0_epoch19_auroc0.9781.ckpt,825944028,59230ec1bae56c934d6c3fe03e6e07eb7db56af6a11dca2cb4adc78d435b6085
+models/remote_true/separation_loss_ablation/sep_loss_lambda_0p0_2026-02-20_16-33-09_sep_0.0_epoch29_auroc0.9648.ckpt,825947740,f8d92995b708316ceacb5c62f2261bb6bbdc6b0ba13ea9b8c4e4a7c7b9c5c2b9
+models/remote_true/separation_loss_ablation/sep_loss_lambda_0p0_2026-02-21_05-04-31_sep_0.0_epoch79_auroc0.8025.ckpt,825947612,ca6cb487d94c1d376d4cf98d74cd5a4491ddefbbbe32c47203cd35f91b93eee6
+models/remote_true/separation_loss_ablation/sep_loss_lambda_0p1_2026-02-22_14-36-51_sep_0.1_epoch149_auroc0.9667.ckpt,825944092,e483f3be193ac7a96b0dd2a8888891b6cac7a67ad9dbe7cf6ff7abeafae8d49b
+models/remote_true/separation_loss_ablation/sep_loss_lambda_0p1_2026-02-24_06-00-19_sep_0.1_epoch09_auroc0.8397.ckpt,825943900,7dcb69acb6fc84cfa54bfe0b501ab42b261d12548fcc064af0eb1888b2c00fef
+models/remote_true/separation_loss_ablation/separation_loss_ablation_models.csv,4393,2e4d6237fbcad74deede88ba0c4d16ad05a7851e1a806020fd0c10fe25804b83
+README.md,1127,eb04102a85991698e1f62619853557640040f980e285fc1fc19e5e7f76761563

missing_checkpoints.md

@@ -0,0 +1,20 @@
+# Missing CIFAR-10 Thesis Checkpoints
+
+The final thesis documentation references checkpoints on old NFS paths that are not present under `D:\side_hustle\thesis\draft02\ssh00`. They are not substituted with similarly named local files.
+
+## Main Training, lambda=0.01
+| Seed | AUROC | Best Epoch | Original Path |
+|---|---:|---:|---|
+| 42 | 0.9873 | 19 | `/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/seed42/2026-02-17_18-18-51_seed42/` |
+| 123 | 0.9886 | 19 | `/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/seed123/2026-02-19_06-55-43_seed123/` |
+| 456 | 0.9887 | 19 | `/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/seed456/2026-02-19_06-56-23_seed456/` |
+
+## Separation-Loss Ablation, seed=42
+| Lambda | AUROC | Best Epoch | Original Path |
+|---:|---:|---:|---|
+| 0.0 | 0.8025 | 79 | `/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/sep_loss_ablation/2026-02-21_05-04-31_sep_0.0/` |
+| 0.001 | 0.9732 | 19 | `/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/sep_loss_ablation/2026-02-21_22-03-43_sep_0.001/` |
+| 0.01 | 0.9869 | 19 | `/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/sep_loss_ablation/2026-02-23_18-52-30_sep_0.01/` |
+| 0.02 | 0.9911 | 29 | `/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/sep_loss_ablation/2026-02-24_22-57-58_sep_0.02/` |
+| 0.05 | 0.9851 | 19 | `/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/sep_loss_ablation/2026-02-23_02-04-16_sep_0.05/` |
+| 0.1 | 0.9667 | 149 | `/system/user/studentwork/mohammed/2025/diffusion_classifier_ood/results/sep_loss_ablation/2026-02-24_06-00-19_sep_0.1/` |

models/raw_scores/seed123_cifar100_scores.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eea96e85415d707344682bc9740402ef7536ad20cdfab4f15d18e4ad7169fbc3
+size 41324

models/raw_scores/seed123_cifar10_id_scores.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f7392a8857171beb7d6c4704d21c0234d11403b8e1436b347c81094a03ecc087
+size 5302

models/raw_scores/seed123_cifar10_ood_scores.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a9a23641d6365e13b0b97aabb2322ab5d5f22e40970530aa3d90256e16be8f25
+size 37307

models/raw_scores/seed123_fashionmnist_scores.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c36a3aaa549d85e0d051bffd39aaa41f3becd1c5a77d256440cc6cec85eca94a
+size 41344

models/raw_scores/seed123_food101_scores.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:efea6c9959a2e1eb8c8f45ec805640e2e9788e3f1335adf2c5f8ffd87456406f
+size 21287

models/raw_scores/seed123_stl10_scores.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:99f782f861fd1967d3114512f3609567d462c82a3bf94fa50b06e84b9fe8ba37
+size 21213

models/raw_scores/seed123_svhn_scores.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:18467a14e7aa7acb41c6867dac363dc8c01d45aa9f8049e4f7e86f3b0af37c43
+size 105368

models/raw_scores/seed123_textures_scores.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3348acbae4bac9d891c40e7f87584ca2003e8185f55a2db31589a11bc7202e6a
+size 8812

models/raw_scores/seed42_cifar100_scores.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:01da976c1ef147ce098120dc72dca4309b039d3deb0582312f19f450a6b9b923
+size 41319

models/raw_scores/seed42_cifar10_id_scores.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f45329e41b35cfb442143379ebdc9d96f488a5f04776ee3d0b4243dec458a757
+size 5297

models/raw_scores/seed42_cifar10_ood_scores.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:64599a919c1dc17d4633ad6024d29d9e6f549fbf491420c9dedae234240e1d39
+size 37302

models/raw_scores/seed42_fashionmnist_scores.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:69bb05b930eaf55bb5e29c8bbe8d93363398fc480be32a497700919a0a488383
+size 41339

models/raw_scores/seed42_food101_scores.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f7c37c628db9b479196923024de79410d3a5bcf15419698657237df36b2beb1b
+size 21218

models/raw_scores/seed42_places365_scores.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fa05e2f8e7c80f1dcf3f056aedf5f319e4fb6b87e54ccbd9c6c10326ddab271c
+size 41324

models/raw_scores/seed42_stl10_scores.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:63bc986fe2d1750424b3d161056b4deb448580a3665c0265070728173142adad
+size 21208

models/raw_scores/seed42_svhn_scores.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1d4b841177293ab8ba149b22d031c8af7e79938dcbaf798d4546c396fffba50f
+size 105363

models/raw_scores/seed42_textures_scores.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8b5a2cf6d56c7c4ee7ee1db6f0a169ec8af01adaa2d65a0fa24430f3dd9f5d38
+size 8807

models/raw_scores/seed456_cifar100_scores.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6f5eff496e89464f01af6dac0a09e9b9ea4c2333cec255654bc9a72c9acc7ba6
+size 41324

models/raw_scores/seed456_cifar10_id_scores.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eca890ba2edb578cd2a1f6549295f4212e7287c180e02e8477f751845f7cd579
+size 5302

models/raw_scores/seed456_cifar10_ood_scores.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3d464ff0c432a15cb6c32fce045b7f2088f91ba6a080ff2d4b27a6bb73dea089
+size 37307

models/raw_scores/seed456_fashionmnist_scores.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:089afd7bc9661058659d07869e2c1400bbce7366df49d4507390d2852fd6692e
+size 41344

models/raw_scores/seed456_food101_scores.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5905ea30ce29afc2b3eea511585a3f5e1cc95d1df3a1150fedada9c17f23e545
+size 21287

models/raw_scores/seed456_stl10_scores.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6ebfbb38187ca20426cbdaa3547f81aa5e4d3830a69c6e623d8a06bfe82fbf8f
+size 21213

models/raw_scores/seed456_svhn_scores.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bbe9b31182ed334606cc8156122ad6f4da2459e90961d7f1c60ae58e271e8107
+size 105368