Archive 4 checkpoints (1-4)

paper-releases/pbisc-diffusion/updates/20261008T022410Z-deconv-scsemi-deck/diffusion/dataset/server_1/model8/plans/paired_finetune_intradataset_20260927/scores/ye2026_sepsis/scsemi_cv_20261002/fold0/README.md

@@ -0,0 +1,73 @@
+# score_ft.py output: ye2026_sepsis, 17 scored donors
+
+Arms: scsemi__raw.  Every arm is scored on the same donors against the same observed reference.
+
+## Inputs and fixed contracts
+- Observed reference: `/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scoring_cache/ye2026_sepsis/reference.npz` (sha bac12da8a241), built from datadir:/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/data/ye2026_sepsis;
+  observed cells sum exactly to the features pseudobulk of every donor. Observed labels: /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scoring_cache/ye2026_sepsis/observed_labels/labels/labels.parquet.
+- Generated cells: `generated_counts` read out once with Poisson(seed 20260916) over the **whole NPZ in row order**
+  (`poisson_readout.poisson_integerise`), then the scored donors' rows. These integer counts are used for labels, embedding,
+  depth and the contrast. Pseudobulk uses the summed `generated_counts` (rates), as `score_generated.py`.
+- Labels: frozen f01common scANVI (`predict_scanvi.py`, CPU, zero query epochs), six-type gate ['T cell', 'B cell', 'Monocyte', 'NK cell', 'DC', 'Other']; query genes =
+  the 1,006-gene annotation list minus genes the dataset's sc assay did not measure in every sample (1006 genes).
+- Scoring genes: 6k panel genes measured by both assays in every sample of the dataset (5975 genes).
+
+## Definitions
+1. **Composition**: fraction of each donor's cells per six-type label (all generated cells; all observed QC cells).
+   `comp_l1` = sum over the six labels of |generated - observed| (immicro_common.l1). `comp_l1_obs_loo_null` = L1 of the
+   donor's observed composition to the mean of the *other scored donors'* observed compositions. `composition_top1` =
+   fraction of donors whose generated composition is closest (L1) to their own observed composition.
+2. **Pseudobulk**: `pb_r` = Pearson r of CPM-log1p(summed generated_counts) vs CPM-log1p(observed pseudobulk) on the scoring
+   genes (evaluate_realbulk.cpm_log1p / paired_r). `pb_top1` = the donor's generated pseudobulk is closest (Pearson) to its own
+   observed pseudobulk among the 17 scored donors (immicro_common.retrieval = evaluate_realbulk.top1 arithmetic).
+3. **Precision / coverage / density** (stage_b_embed.py functions): per-cell CP10K over the scoring genes + log1p; PCA-50
+   (randomized, random_state 20260925) and k = 10 real-real k-NN balls fitted on **R1 of the scored donors only**.
+   Reference rule (independent of fold and arm): per donor, cells in canonical order (library_id+barcode), one
+   permutation from `default_rng([20260925, uint64(sha256(donor)[:8])])`; R1 = first n, R2 = next n,
+   n = min(150, n_cells // 2) (2550 R1 cells here). Generated query: per donor n cells (the donor's R1 size),
+   without replacement, `default_rng([20260926, same key])`. Precision = fraction of generated cells inside any R1 ball
+   (Kynkaanniemi 2019); coverage = fraction of R1 cells whose ball holds a generated cell; density = Naeem 2020.
+   Per-donor precision = that donor's generated cells; per-donor coverage = that donor's R1 cells (the pooled values are
+   their n-weighted means; plain means when every donor has 150 reference cells). R2 (held-out real cells of the same donors) is the
+   perfect-generator control: precision 0.9969, coverage 1.0000.
+   **Depth matched**: stage_b_embed depth matching run pairwise (R1 vs this arm): per donor and within-donor rank of detected
+   genes, both cells are thinned by UMI sampling without replacement to the smaller detected-gene count; PCA kept from
+   un-thinned R1, radii recomputed on thinned R1; fresh rng(20260929) per arm. R2 control: precision
+   0.9961, coverage 1.0000.
+4. **Monocyte %** = 100 x Monocyte fraction (all cells) per donor; **median UMI per cell** = median over cells of the
+   Poisson-read (generated) / observed counts summed over the scoring genes (per donor, and pooled).
+5. **Group contrast** (T, B, Monocyte, NK): case groups ['adult_sepsis', 'ped_sepsis'] (n = 12) vs control groups ['adult_HC', 'adult_ICU_ctrl', 'ped_HC', 'ped_ICU_ctrl'] (n = 5); rule: explicit --case-groups / --control-groups; groups from features groups.
+   Genes per type = scoring genes detected (count > 0) in >= 1 % of the scored donors' observed
+   cells of that type. Primary statistic `mean_cp10k`: per-cell CP10K over the scoring genes; per donor the mean over its
+   cells of the type (donors with >= 1 such cell); group mean = mean over the group's donors (donor-balanced);
+   logFC = log2(case mean + 1) - log2(control mean + 1). Secondary `pseudobulk_log1p_cpm`: the SLE benchmark evaluator's
+   statistic (donor x type summed counts, log1p CPM over the scoring genes, mean case - mean control;
+   evaluate_generated_vs_truth.group_log_fold_change). Both are computed on the generated cells (their frozen labels) and on
+   the observed cells; `r_generated` = Pearson r of the two logFC vectors. Broadcast baseline `r_broadcast`: the same
+   observed vector against the logFC of the input bulk (condition NPZ `bulk_counts` of the arm, scoring genes, per-donor
+   CP10K resp. log1p CPM, group means). `partial_r_generated_given_broadcast`: r of the OLS residuals of observed and
+   generated logFC on the bulk logFC (residual_specificity.py).
+6. **Top-50 DEG overlap** (contrast.csv rows `statistic = welch_top50_deg`; summary `deg_top50`): the SLE benchmark's score (`total_top50_overlap` of
+   `bulk2sc-paper/scripts/evaluate_generated_vs_truth.py`, Welch backend: the "top-50 DEG 81 / 84" of the final model), run
+   with the evaluator's own functions on the same donor x type pseudobulks as the contrast. Per cell type, observed and
+   generated separately: donor x type summed counts on all 5975 scoring genes (donors of the two groups with
+   >= 1 cell of the type) -> `welch_bh_deg` (log1p CPM, two-sided Welch t-test case vs control, Benjamini-Hochberg over the
+   genes with a defined p-value; needs >= 2 donors per group) -> `top_deg_genes` (FDR < 0.05, descending |logFC|,
+   first 50; shorter when fewer genes pass). `top50_overlap` = |observed list & generated list|; `top50_jaccard`
+   the evaluator's Jaccard. Summary `sum` = total over the scored types (the quantity reported as 81), `max` = the sum of
+   the observed list lengths (the "of 262"), `sum_4type` = T + B + Monocyte + NK. A type is scored when its observed test
+   is defined; an arm whose generated test is undefined for a scored type gets overlap 0. Types: T cell, B cell, Monocyte, NK cell, DC
+   (aggregates: /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scoring_cache/ye2026_sepsis/deg_reference.npz).
+   **Broadcast analogue** (`top50_overlap_broadcast`, as `sle_broadcast.py` `broadcast_input_bulk`; not part of the
+   published benchmark): the same test on the arm's condition bulk_counts (scoring genes, the 12 + 5
+   donors), one whole-PBMC top-50 list, compared with every type's observed list.
+
+## Files
+- `per_donor.csv`: one row per arm x donor (all per-donor quantities above).
+- `summary.json`: per arm the donor means / pooled values, provenance (NPZ sha, receipt check, generation metadata,
+  Poisson nnz of the whole NPZ, annotation contract), and the reference block (PCA, R2 control).
+- `contrast.csv`: per arm x cell type x statistic (`mean_cp10k`, `pseudobulk_log1p_cpm`: logFC r; `welch_top50_deg`:
+  top-50 DEG overlap, T / B / Monocyte / NK / DC; columns of the other kind are empty).
+- `arms/<arm>/labels/` (predict_scanvi.py labels + manifest), `arms/<arm>/generated_cells.parquet` (NPZ row, donor, label,
+  UMI on the scoring genes, P/C subsample flag), `arms/<arm>/deg_top50_lists.csv` (observed, generated and broadcast
+  top-50 gene lists, Ensembl ids, rank order).

paper-releases/pbisc-diffusion/updates/20261008T022410Z-deconv-scsemi-deck/diffusion/dataset/server_1/model8/plans/paired_finetune_intradataset_20260927/scores/ye2026_sepsis/scsemi_cv_20261002/fold1/README.md

@@ -0,0 +1,73 @@
+# score_ft.py output: ye2026_sepsis, 17 scored donors
+
+Arms: scsemi__raw.  Every arm is scored on the same donors against the same observed reference.
+
+## Inputs and fixed contracts
+- Observed reference: `/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scoring_cache/ye2026_sepsis/reference.npz` (sha bac12da8a241), built from datadir:/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/data/ye2026_sepsis;
+  observed cells sum exactly to the features pseudobulk of every donor. Observed labels: /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scoring_cache/ye2026_sepsis/observed_labels/labels/labels.parquet.
+- Generated cells: `generated_counts` read out once with Poisson(seed 20260916) over the **whole NPZ in row order**
+  (`poisson_readout.poisson_integerise`), then the scored donors' rows. These integer counts are used for labels, embedding,
+  depth and the contrast. Pseudobulk uses the summed `generated_counts` (rates), as `score_generated.py`.
+- Labels: frozen f01common scANVI (`predict_scanvi.py`, CPU, zero query epochs), six-type gate ['T cell', 'B cell', 'Monocyte', 'NK cell', 'DC', 'Other']; query genes =
+  the 1,006-gene annotation list minus genes the dataset's sc assay did not measure in every sample (1006 genes).
+- Scoring genes: 6k panel genes measured by both assays in every sample of the dataset (5975 genes).
+
+## Definitions
+1. **Composition**: fraction of each donor's cells per six-type label (all generated cells; all observed QC cells).
+   `comp_l1` = sum over the six labels of |generated - observed| (immicro_common.l1). `comp_l1_obs_loo_null` = L1 of the
+   donor's observed composition to the mean of the *other scored donors'* observed compositions. `composition_top1` =
+   fraction of donors whose generated composition is closest (L1) to their own observed composition.
+2. **Pseudobulk**: `pb_r` = Pearson r of CPM-log1p(summed generated_counts) vs CPM-log1p(observed pseudobulk) on the scoring
+   genes (evaluate_realbulk.cpm_log1p / paired_r). `pb_top1` = the donor's generated pseudobulk is closest (Pearson) to its own
+   observed pseudobulk among the 17 scored donors (immicro_common.retrieval = evaluate_realbulk.top1 arithmetic).
+3. **Precision / coverage / density** (stage_b_embed.py functions): per-cell CP10K over the scoring genes + log1p; PCA-50
+   (randomized, random_state 20260925) and k = 10 real-real k-NN balls fitted on **R1 of the scored donors only**.
+   Reference rule (independent of fold and arm): per donor, cells in canonical order (library_id+barcode), one
+   permutation from `default_rng([20260925, uint64(sha256(donor)[:8])])`; R1 = first n, R2 = next n,
+   n = min(150, n_cells // 2) (2550 R1 cells here). Generated query: per donor n cells (the donor's R1 size),
+   without replacement, `default_rng([20260926, same key])`. Precision = fraction of generated cells inside any R1 ball
+   (Kynkaanniemi 2019); coverage = fraction of R1 cells whose ball holds a generated cell; density = Naeem 2020.
+   Per-donor precision = that donor's generated cells; per-donor coverage = that donor's R1 cells (the pooled values are
+   their n-weighted means; plain means when every donor has 150 reference cells). R2 (held-out real cells of the same donors) is the
+   perfect-generator control: precision 0.9980, coverage 1.0000.
+   **Depth matched**: stage_b_embed depth matching run pairwise (R1 vs this arm): per donor and within-donor rank of detected
+   genes, both cells are thinned by UMI sampling without replacement to the smaller detected-gene count; PCA kept from
+   un-thinned R1, radii recomputed on thinned R1; fresh rng(20260929) per arm. R2 control: precision
+   0.9965, coverage 1.0000.
+4. **Monocyte %** = 100 x Monocyte fraction (all cells) per donor; **median UMI per cell** = median over cells of the
+   Poisson-read (generated) / observed counts summed over the scoring genes (per donor, and pooled).
+5. **Group contrast** (T, B, Monocyte, NK): case groups ['adult_sepsis', 'ped_sepsis'] (n = 12) vs control groups ['adult_HC', 'adult_ICU_ctrl', 'ped_HC', 'ped_ICU_ctrl'] (n = 5); rule: explicit --case-groups / --control-groups; groups from features groups.
+   Genes per type = scoring genes detected (count > 0) in >= 1 % of the scored donors' observed
+   cells of that type. Primary statistic `mean_cp10k`: per-cell CP10K over the scoring genes; per donor the mean over its
+   cells of the type (donors with >= 1 such cell); group mean = mean over the group's donors (donor-balanced);
+   logFC = log2(case mean + 1) - log2(control mean + 1). Secondary `pseudobulk_log1p_cpm`: the SLE benchmark evaluator's
+   statistic (donor x type summed counts, log1p CPM over the scoring genes, mean case - mean control;
+   evaluate_generated_vs_truth.group_log_fold_change). Both are computed on the generated cells (their frozen labels) and on
+   the observed cells; `r_generated` = Pearson r of the two logFC vectors. Broadcast baseline `r_broadcast`: the same
+   observed vector against the logFC of the input bulk (condition NPZ `bulk_counts` of the arm, scoring genes, per-donor
+   CP10K resp. log1p CPM, group means). `partial_r_generated_given_broadcast`: r of the OLS residuals of observed and
+   generated logFC on the bulk logFC (residual_specificity.py).
+6. **Top-50 DEG overlap** (contrast.csv rows `statistic = welch_top50_deg`; summary `deg_top50`): the SLE benchmark's score (`total_top50_overlap` of
+   `bulk2sc-paper/scripts/evaluate_generated_vs_truth.py`, Welch backend: the "top-50 DEG 81 / 84" of the final model), run
+   with the evaluator's own functions on the same donor x type pseudobulks as the contrast. Per cell type, observed and
+   generated separately: donor x type summed counts on all 5975 scoring genes (donors of the two groups with
+   >= 1 cell of the type) -> `welch_bh_deg` (log1p CPM, two-sided Welch t-test case vs control, Benjamini-Hochberg over the
+   genes with a defined p-value; needs >= 2 donors per group) -> `top_deg_genes` (FDR < 0.05, descending |logFC|,
+   first 50; shorter when fewer genes pass). `top50_overlap` = |observed list & generated list|; `top50_jaccard`
+   the evaluator's Jaccard. Summary `sum` = total over the scored types (the quantity reported as 81), `max` = the sum of
+   the observed list lengths (the "of 262"), `sum_4type` = T + B + Monocyte + NK. A type is scored when its observed test
+   is defined; an arm whose generated test is undefined for a scored type gets overlap 0. Types: T cell, B cell, Monocyte, NK cell, DC
+   (aggregates: /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scoring_cache/ye2026_sepsis/deg_reference.npz).
+   **Broadcast analogue** (`top50_overlap_broadcast`, as `sle_broadcast.py` `broadcast_input_bulk`; not part of the
+   published benchmark): the same test on the arm's condition bulk_counts (scoring genes, the 12 + 5
+   donors), one whole-PBMC top-50 list, compared with every type's observed list.
+
+## Files
+- `per_donor.csv`: one row per arm x donor (all per-donor quantities above).
+- `summary.json`: per arm the donor means / pooled values, provenance (NPZ sha, receipt check, generation metadata,
+  Poisson nnz of the whole NPZ, annotation contract), and the reference block (PCA, R2 control).
+- `contrast.csv`: per arm x cell type x statistic (`mean_cp10k`, `pseudobulk_log1p_cpm`: logFC r; `welch_top50_deg`:
+  top-50 DEG overlap, T / B / Monocyte / NK / DC; columns of the other kind are empty).
+- `arms/<arm>/labels/` (predict_scanvi.py labels + manifest), `arms/<arm>/generated_cells.parquet` (NPZ row, donor, label,
+  UMI on the scoring genes, P/C subsample flag), `arms/<arm>/deg_top50_lists.csv` (observed, generated and broadcast
+  top-50 gene lists, Ensembl ids, rank order).

paper-releases/pbisc-diffusion/updates/20261008T022410Z-deconv-scsemi-deck/diffusion/dataset/server_1/model8/plans/paired_finetune_intradataset_20260927/scores/ye2026_sepsis/scsemi_cv_20261002/fold2/README.md

@@ -0,0 +1,73 @@
+# score_ft.py output: ye2026_sepsis, 16 scored donors
+
+Arms: scsemi__raw.  Every arm is scored on the same donors against the same observed reference.
+
+## Inputs and fixed contracts
+- Observed reference: `/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scoring_cache/ye2026_sepsis/reference.npz` (sha bac12da8a241), built from datadir:/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/data/ye2026_sepsis;
+  observed cells sum exactly to the features pseudobulk of every donor. Observed labels: /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scoring_cache/ye2026_sepsis/observed_labels/labels/labels.parquet.
+- Generated cells: `generated_counts` read out once with Poisson(seed 20260916) over the **whole NPZ in row order**
+  (`poisson_readout.poisson_integerise`), then the scored donors' rows. These integer counts are used for labels, embedding,
+  depth and the contrast. Pseudobulk uses the summed `generated_counts` (rates), as `score_generated.py`.
+- Labels: frozen f01common scANVI (`predict_scanvi.py`, CPU, zero query epochs), six-type gate ['T cell', 'B cell', 'Monocyte', 'NK cell', 'DC', 'Other']; query genes =
+  the 1,006-gene annotation list minus genes the dataset's sc assay did not measure in every sample (1006 genes).
+- Scoring genes: 6k panel genes measured by both assays in every sample of the dataset (5975 genes).
+
+## Definitions
+1. **Composition**: fraction of each donor's cells per six-type label (all generated cells; all observed QC cells).
+   `comp_l1` = sum over the six labels of |generated - observed| (immicro_common.l1). `comp_l1_obs_loo_null` = L1 of the
+   donor's observed composition to the mean of the *other scored donors'* observed compositions. `composition_top1` =
+   fraction of donors whose generated composition is closest (L1) to their own observed composition.
+2. **Pseudobulk**: `pb_r` = Pearson r of CPM-log1p(summed generated_counts) vs CPM-log1p(observed pseudobulk) on the scoring
+   genes (evaluate_realbulk.cpm_log1p / paired_r). `pb_top1` = the donor's generated pseudobulk is closest (Pearson) to its own
+   observed pseudobulk among the 16 scored donors (immicro_common.retrieval = evaluate_realbulk.top1 arithmetic).
+3. **Precision / coverage / density** (stage_b_embed.py functions): per-cell CP10K over the scoring genes + log1p; PCA-50
+   (randomized, random_state 20260925) and k = 10 real-real k-NN balls fitted on **R1 of the scored donors only**.
+   Reference rule (independent of fold and arm): per donor, cells in canonical order (library_id+barcode), one
+   permutation from `default_rng([20260925, uint64(sha256(donor)[:8])])`; R1 = first n, R2 = next n,
+   n = min(150, n_cells // 2) (2400 R1 cells here). Generated query: per donor n cells (the donor's R1 size),
+   without replacement, `default_rng([20260926, same key])`. Precision = fraction of generated cells inside any R1 ball
+   (Kynkaanniemi 2019); coverage = fraction of R1 cells whose ball holds a generated cell; density = Naeem 2020.
+   Per-donor precision = that donor's generated cells; per-donor coverage = that donor's R1 cells (the pooled values are
+   their n-weighted means; plain means when every donor has 150 reference cells). R2 (held-out real cells of the same donors) is the
+   perfect-generator control: precision 0.9971, coverage 1.0000.
+   **Depth matched**: stage_b_embed depth matching run pairwise (R1 vs this arm): per donor and within-donor rank of detected
+   genes, both cells are thinned by UMI sampling without replacement to the smaller detected-gene count; PCA kept from
+   un-thinned R1, radii recomputed on thinned R1; fresh rng(20260929) per arm. R2 control: precision
+   0.9962, coverage 1.0000.
+4. **Monocyte %** = 100 x Monocyte fraction (all cells) per donor; **median UMI per cell** = median over cells of the
+   Poisson-read (generated) / observed counts summed over the scoring genes (per donor, and pooled).
+5. **Group contrast** (T, B, Monocyte, NK): case groups ['adult_sepsis', 'ped_sepsis'] (n = 12) vs control groups ['adult_HC', 'adult_ICU_ctrl', 'ped_HC', 'ped_ICU_ctrl'] (n = 4); rule: explicit --case-groups / --control-groups; groups from features groups.
+   Genes per type = scoring genes detected (count > 0) in >= 1 % of the scored donors' observed
+   cells of that type. Primary statistic `mean_cp10k`: per-cell CP10K over the scoring genes; per donor the mean over its
+   cells of the type (donors with >= 1 such cell); group mean = mean over the group's donors (donor-balanced);
+   logFC = log2(case mean + 1) - log2(control mean + 1). Secondary `pseudobulk_log1p_cpm`: the SLE benchmark evaluator's
+   statistic (donor x type summed counts, log1p CPM over the scoring genes, mean case - mean control;
+   evaluate_generated_vs_truth.group_log_fold_change). Both are computed on the generated cells (their frozen labels) and on
+   the observed cells; `r_generated` = Pearson r of the two logFC vectors. Broadcast baseline `r_broadcast`: the same
+   observed vector against the logFC of the input bulk (condition NPZ `bulk_counts` of the arm, scoring genes, per-donor
+   CP10K resp. log1p CPM, group means). `partial_r_generated_given_broadcast`: r of the OLS residuals of observed and
+   generated logFC on the bulk logFC (residual_specificity.py).
+6. **Top-50 DEG overlap** (contrast.csv rows `statistic = welch_top50_deg`; summary `deg_top50`): the SLE benchmark's score (`total_top50_overlap` of
+   `bulk2sc-paper/scripts/evaluate_generated_vs_truth.py`, Welch backend: the "top-50 DEG 81 / 84" of the final model), run
+   with the evaluator's own functions on the same donor x type pseudobulks as the contrast. Per cell type, observed and
+   generated separately: donor x type summed counts on all 5975 scoring genes (donors of the two groups with
+   >= 1 cell of the type) -> `welch_bh_deg` (log1p CPM, two-sided Welch t-test case vs control, Benjamini-Hochberg over the
+   genes with a defined p-value; needs >= 2 donors per group) -> `top_deg_genes` (FDR < 0.05, descending |logFC|,
+   first 50; shorter when fewer genes pass). `top50_overlap` = |observed list & generated list|; `top50_jaccard`
+   the evaluator's Jaccard. Summary `sum` = total over the scored types (the quantity reported as 81), `max` = the sum of
+   the observed list lengths (the "of 262"), `sum_4type` = T + B + Monocyte + NK. A type is scored when its observed test
+   is defined; an arm whose generated test is undefined for a scored type gets overlap 0. Types: T cell, B cell, Monocyte, NK cell, DC
+   (aggregates: /home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scoring_cache/ye2026_sepsis/deg_reference.npz).
+   **Broadcast analogue** (`top50_overlap_broadcast`, as `sle_broadcast.py` `broadcast_input_bulk`; not part of the
+   published benchmark): the same test on the arm's condition bulk_counts (scoring genes, the 12 + 4
+   donors), one whole-PBMC top-50 list, compared with every type's observed list.
+
+## Files
+- `per_donor.csv`: one row per arm x donor (all per-donor quantities above).
+- `summary.json`: per arm the donor means / pooled values, provenance (NPZ sha, receipt check, generation metadata,
+  Poisson nnz of the whole NPZ, annotation contract), and the reference block (PCA, R2 control).
+- `contrast.csv`: per arm x cell type x statistic (`mean_cp10k`, `pseudobulk_log1p_cpm`: logFC r; `welch_top50_deg`:
+  top-50 DEG overlap, T / B / Monocyte / NK / DC; columns of the other kind are empty).
+- `arms/<arm>/labels/` (predict_scanvi.py labels + manifest), `arms/<arm>/generated_cells.parquet` (NPZ row, donor, label,
+  UMI on the scoring genes, P/C subsample flag), `arms/<arm>/deg_top50_lists.csv` (observed, generated and broadcast
+  top-50 gene lists, Ensembl ids, rank order).

paper-releases/pbisc-diffusion/updates/20261008T022410Z-deconv-scsemi-deck/diffusion/dataset/server_1/model8/plans/paired_finetune_intradataset_20260927/scores/ye2026_sepsis/scsemi_cv_20261002/pooled/summary.json

@@ -0,0 +1,765 @@
+{
+ "schema": "paired_ft_scores.v1",
+ "dataset": "ye2026_sepsis",
+ "written_utc": "2026-10-07T03:02:34.573934+00:00",
+ "command": [
+  "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/tools/score_ft.py",
+  "--dataset",
+  "ye2026_sepsis",
+  "--out",
+  "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scores/ye2026_sepsis/scsemi_cv_20261002/pooled",
+  "--threads",
+  "8",
+  "--arm",
+  "scsemi__raw=/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/generated/ye2026_sepsis/pooled_scsemi_cv_20261002/scsemi__raw.npz",
+  "--case-groups",
+  "adult_sepsis,ped_sepsis",
+  "--control-groups",
+  "adult_HC,adult_ICU_ctrl,ped_HC,ped_ICU_ctrl"
+ ],
+ "arms": [
+  "scsemi__raw"
+ ],
+ "donors": [
+  "Abd-PS04",
+  "Abd-PS05",
+  "Abd-PS07",
+  "Abd-PS19",
+  "Abd-PS20",
+  "Abd-PS23",
+  "Abd-PS27",
+  "Abd-S102",
+  "Abd-S103",
+  "Abd-S120",
+  "Abd-S123",
+  "Abd-S156",
+  "Abd-S39",
+  "Abd-S42",
+  "Abd-S47",
+  "Abd-S51",
+  "Abd-S54",
+  "Abd-S57",
+  "Abd-S66",
+  "Abd-S79",
+  "Abd-S83",
+  "Abd-S98",
+  "HC01",
+  "HC05",
+  "HC07",
+  "HC09",
+  "HC12",
+  "HC13",
+  "HC16",
+  "IC10",
+  "IC15",
+  "IC16",
+  "IC19",
+  "IC22",
+  "IC23",
+  "PHC01",
+  "PHC02",
+  "PHC03",
+  "PHC05",
+  "PIC01",
+  "PIC02",
+  "PIC04",
+  "PIC05",
+  "Res-PS02",
+  "Res-PS03",
+  "Res-PS12",
+  "Res-PS16",
+  "Res-PS21",
+  "Res-PS22",
+  "Res-PS25",
+  "Res-PS26",
+  "Res-S10",
+  "Res-S125",
+  "Res-S127",
+  "Res-S136",
+  "Res-S139",
+  "Res-S146",
+  "Res-S149",
+  "Res-S150",
+  "Res-S164",
+  "Res-S166",
+  "Res-S174",
+  "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"
+ ],
+ "n_donors": 82,
+ "reference": {
+  "cache": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scoring_cache/ye2026_sepsis",
+  "reference_sha256": "bac12da8a241652cb75100a1b082b2be56806c9bf2fd20ce6f927f19683ea5b8",
+  "features_sha256": "0cfda4b0cf1caf5e5319d071697c228d894e992fd090e2b3ed450bd6c9887d46",
+  "n_r1_cells": 12300,
+  "pca": {
+   "n_components": 50,
+   "explained_variance_ratio_sum": 0.33549821376800537,
+   "n_fit_cells": 12300
+  },
+  "r2_control": {
+   "precision": 0.9804878048780488,
+   "coverage": 0.9996747967479674,
+   "density": 1.2890731707317074,
+   "n_query": 12300,
+   "depth_matched": {
+    "precision": 0.9795934959349594,
+    "coverage": 0.9998373983739838,
+    "density": 1.2835853658536585
+   }
+  },
+  "observed_l1_loo_null_mean": 0.515468004498372
+ },
+ "contrast_definition": {
+  "enabled": true,
+  "rule": "explicit --case-groups / --control-groups",
+  "group_source": "features groups",
+  "case_groups": [
+   "adult_sepsis",
+   "ped_sepsis"
+  ],
+  "control_groups": [
+   "adult_HC",
+   "adult_ICU_ctrl",
+   "ped_HC",
+   "ped_ICU_ctrl"
+  ],
+  "n_case_donors": 61,
+  "n_control_donors": 21,
+  "n_donors_in_neither": 0,
+  "case": [
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true
+  ],
+  "ctrl": [
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   true,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false,
+   false
+  ],
+  "groups": [
+   "ped_sepsis",
+   "ped_sepsis",
+   "ped_sepsis",
+   "ped_sepsis",
+   "ped_sepsis",
+   "ped_sepsis",
+   "ped_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_HC",
+   "adult_HC",
+   "adult_HC",
+   "adult_HC",
+   "adult_HC",
+   "adult_HC",
+   "adult_HC",
+   "adult_ICU_ctrl",
+   "adult_ICU_ctrl",
+   "adult_ICU_ctrl",
+   "adult_ICU_ctrl",
+   "adult_ICU_ctrl",
+   "adult_ICU_ctrl",
+   "ped_HC",
+   "ped_HC",
+   "ped_HC",
+   "ped_HC",
+   "ped_ICU_ctrl",
+   "ped_ICU_ctrl",
+   "ped_ICU_ctrl",
+   "ped_ICU_ctrl",
+   "ped_sepsis",
+   "ped_sepsis",
+   "ped_sepsis",
+   "ped_sepsis",
+   "ped_sepsis",
+   "ped_sepsis",
+   "ped_sepsis",
+   "ped_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis",
+   "adult_sepsis"
+  ]
+ },
+ "deg_definition": {
+  "enabled": true,
+  "score": "top-50 DEG overlap per cell type (benchmark total_top50_overlap, Welch backend)",
+  "source": "/home/sj_server_1/PBISC/immun-pbmc-fm/bulk2sc-paper/scripts/evaluate_generated_vs_truth.py",
+  "functions": [
+   "welch_bh_deg",
+   "top_deg_genes",
+   "jaccard"
+  ],
+  "reproduces": "utility_studies_final6k_20260925/sle top-50 DEG 81 (1,006 genes) / 84 (2,279 genes) = sum over the six gated types of top50_overlap_count",
+  "top_n": 50,
+  "fdr": 0.05,
+  "n_genes": 5975,
+  "types": [
+   "T cell",
+   "B cell",
+   "Monocyte",
+   "NK cell",
+   "DC"
+  ],
+  "aggregates_source": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/scoring_cache/ye2026_sepsis/deg_reference.npz",
+  "aggregates_sha256": "9886ea64da90b1db6a07984d5d61d7dab04ee0beee7d4094ee3672316c2691d0",
+  "aggregates_note": null,
+  "types_not_scored_observed": [],
+  "observed": {
+   "T cell": {
+    "n_significant": 1578,
+    "n_tested": 5810,
+    "n_case_donors": 61,
+    "n_control_donors": 21,
+    "n_cells": 298117,
+    "top_list_size": 50
+   },
+   "B cell": {
+    "n_significant": 1102,
+    "n_tested": 5752,
+    "n_case_donors": 61,
+    "n_control_donors": 21,
+    "n_cells": 99729,
+    "top_list_size": 50
+   },
+   "Monocyte": {
+    "n_significant": 2459,
+    "n_tested": 5754,
+    "n_case_donors": 61,
+    "n_control_donors": 21,
+    "n_cells": 111269,
+    "top_list_size": 50
+   },
+   "NK cell": {
+    "n_significant": 1235,
+    "n_tested": 5623,
+    "n_case_donors": 61,
+    "n_control_donors": 21,
+    "n_cells": 47832,
+    "top_list_size": 50
+   },
+   "DC": {
+    "n_significant": 1572,
+    "n_tested": 5330,
+    "n_case_donors": 57,
+    "n_control_donors": 21,
+    "n_cells": 3300,
+    "top_list_size": 50
+   }
+  },
+  "broadcast": "welch_bh_deg + top_deg_genes on the arm's condition bulk_counts (scoring genes, case|control donors), one list compared with every type's observed list (sle_broadcast.py broadcast_input_bulk); not part of the published benchmark"
+ },
+ "headline": {
+  "scsemi__raw": {
+   "comp_l1": 0.5022299158951089,
+   "comp_l1_obs_loo_null": 0.515468004498372,
+   "precision": 0.7596747967479675,
+   "coverage": 0.09601626016260163,
+   "precision_dm": 0.5473170731707317,
+   "coverage_dm": 0.20991869918699188,
+   "pb_r_median": 0.6129241550465101,
+   "pb_top1": 0.012195121951219513,
+   "mono_pct_gen": 22.234146341463415,
+   "mono_pct_obs": 20.817070783198325,
+   "median_umi_gen": 1967.0,
+   "median_umi_obs": 1967.0,
+   "contrast_mean_cp10k_r_gen": 0.43526410070392063,
+   "contrast_pseudobulk_log1p_cpm_r_gen": 0.21734344370825287,
+   "contrast_mean_cp10k_r_broadcast": 0.31594581626202867,
+   "contrast_pseudobulk_log1p_cpm_r_broadcast": 0.29096860411544245,
+   "deg_top50_T": 1,
+   "deg_top50_B": 1,
+   "deg_top50_Mono": 3,
+   "deg_top50_NK": 0,
+   "deg_top50_DC": 1,
+   "deg_top50_sum": 6,
+   "deg_top50_max": 250,
+   "deg_top50_sum_4type": 5,
+   "deg_top50_sum_broadcast": 21,
+   "seconds": 29.0
+  }
+ },
+ "per_arm": {
+  "scsemi__raw": {
+   "arm": "scsemi__raw",
+   "npz": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/generated/ye2026_sepsis/pooled_scsemi_cv_20261002/scsemi__raw.npz",
+   "npz_sha256": "9eb870cbc6b56135fd5c575b0939529141c8ffca90751934ab0350b9cfe23a49",
+   "receipt_verified": true,
+   "generation": {
+    "checkpoint": null,
+    "checkpoint_sha256": null,
+    "input_condition_path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/generated/ye2026_sepsis/pooled_scsemi_cv_20261002/conditions_pooled_raw.npz",
+    "input_condition_sha256": "3b2d8ef2a51603a490100c336f3702cdac11d1ba2288490e36dd52986f56dd13",
+    "seed": null,
+    "steps": null,
+    "support_readout": null,
+    "weights": null
+   },
+   "condition_for_broadcast": {
+    "path": "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/generated/ye2026_sepsis/pooled_scsemi_cv_20261002/conditions_pooled_raw.npz",
+    "sha256": "3b2d8ef2a51603a490100c336f3702cdac11d1ba2288490e36dd52986f56dd13"
+   },
+   "poisson": {
+    "seed": 20260916,
+    "rule": "whole NPZ, rows in NPZ order",
+    "nnz_whole_npz": 14491115,
+    "median_library_whole_npz_panel": 1967.0,
+    "n_rows_whole_npz": 123000
+   },
+   "annotation": {
+    "model_pt_sha256": "bf3655899e5abee435340fc008656c4a17342da0f30b70cfd43de8e9223e1930",
+    "zero_padded_model_genes": 0,
+    "label_counts": {
+     "T cell": 65670,
+     "Monocyte": 27348,
+     "B cell": 14754,
+     "Other": 6569,
+     "NK cell": 6290,
+     "DC": 2369
+    },
+    "mean_l2_max_prob": 0.9213427455382861,
+    "predict_seconds": 8.0
+   },
+   "n_donors": 82,
+   "n_generated_cells": 123000,
+   "composition": {
+    "l1_mean": 0.5022299158951089,
+    "l1_median": 0.41869888957937296,
+    "l1_obs_loo_null_mean": 0.515468004498372,
+    "l1_generated_loo_mean_vs_observed_mean": 0.5159913792704983,
+    "composition_top1": 0.04878048780487805,
+    "mean_fraction_generated": {
+     "T cell": 0.533902,
+     "B cell": 0.119951,
+     "Monocyte": 0.222341,
+     "NK cell": 0.051138,
+     "DC": 0.01926,
+     "Other": 0.053407
+    },
+    "mean_fraction_observed": {
+     "T cell": 0.484547,
+     "B cell": 0.172757,
+     "Monocyte": 0.208171,
+     "NK cell": 0.080723,
+     "DC": 0.005542,
+     "Other": 0.04826
+    }
+   },
+   "pseudobulk": {
+    "n_genes": 5975,
+    "r_median": 0.6129241550465101,
+    "r_mean": 0.6082355090759471,
+    "top1": 0.012195121951219513,
+    "n_top1": 1,
+    "own_rank_median": 18.5,
+    "chance_top1": 0.012195121951219513
+   },
+   "precision_coverage": {
+    "precision": 0.7596747967479675,
+    "coverage": 0.09601626016260163,
+    "density": 0.22936585365853657,
+    "n_query": 12300,
+    "n_reference": 12300
+   },
+   "precision_coverage_depth_matched": {
+    "precision": 0.5473170731707317,
+    "coverage": 0.20991869918699188,
+    "density": 0.20260162601626014,
+    "detected_median_before": {
+     "R1": 411.0,
+     "Q": 116.0
+    },
+    "detected_median_after": {
+     "R1": 116.0,
+     "Q": 116.0
+    }
+   },
+   "monocyte_pct": {
+    "generated_mean_over_donors": 22.234146341463415,
+    "observed_mean_over_donors": 20.817070783198325,
+    "generated_pooled_cells": 22.234146341463415
+   },
+   "median_umi_per_cell_scoring_genes": {
+    "generated_pooled": 1967.0,
+    "observed_pooled": 1967.0,
+    "generated_mean_of_donor_medians": 1967.1158447265625,
+    "observed_mean_of_donor_medians": 2112.9146341463415
+   },
+   "contrast": {
+    "mean_cp10k": {
+     "per_type": {
+      "T cell": {
+       "r_generated": 0.37777476829221296,
+       "r_broadcast": 0.2548893770824626,
+       "partial_r_generated_given_broadcast": 0.30450666006975935,
+       "r_generated_vs_broadcast": 0.4511601716530676,
+       "n_genes": 3258
+      },
+      "B cell": {
+       "r_generated": 0.32571976493328914,
+       "r_broadcast": 0.27458639801653467,
+       "partial_r_generated_given_broadcast": 0.2705465870269831,
+       "r_generated_vs_broadcast": 0.27545603380804967,
+       "n_genes": 3310
+      },
+      "Monocyte": {
+       "r_generated": 0.5835203193635443,
+       "r_broadcast": 0.5481103923782986,
+       "partial_r_generated_given_broadcast": 0.4358649661929322,
+       "r_generated_vs_broadcast": 0.4817528015996326,
+       "n_genes": 3289
+      },
+      "NK cell": {
+       "r_generated": 0.4540415502266361,
+       "r_broadcast": 0.1861970975708186,
+       "partial_r_generated_given_broadcast": 0.4218717919988914,
+       "r_generated_vs_broadcast": 0.44542475698561723,
+       "n_genes": 3240
+      }
+     },
+     "mean_r_generated": 0.43526410070392063,
+     "mean_r_broadcast": 0.31594581626202867,
+     "mean_partial_r": 0.3581975013221415
+    },
+    "pseudobulk_log1p_cpm": {
+     "per_type": {
+      "T cell": {
+       "r_generated": 0.2317567821636163,
+       "r_broadcast": 0.2164745523777111,
+       "partial_r_generated_given_broadcast": 0.19371787584333694,
+       "r_generated_vs_broadcast": 0.21793984219477572,
+       "n_genes": 3258
+      },
+      "B cell": {
+       "r_generated": 0.2419264255614331,
+       "r_broadcast": 0.2711145517540278,
+       "partial_r_generated_given_broadcast": 0.20808010290928658,
+       "r_generated_vs_broadcast": 0.16353149301776773,
+       "n_genes": 3310
+      },
+      "Monocyte": {
+       "r_generated": 0.19462942906730224,
+       "r_broadcast": 0.6203629305570988,
+       "partial_r_generated_given_broadcast": -0.025438390460109463,
+       "r_generated_vs_broadcast": 0.34393408878073184,
+       "n_genes": 3289
+      },
+      "NK cell": {
+       "r_generated": 0.2010611380406599,
+       "r_broadcast": 0.05592238177293225,
+       "partial_r_generated_given_broadcast": 0.19544261871381305,
+       "r_generated_vs_broadcast": 0.14060253076107798,
+       "n_genes": 3240
+      }
+     },
+     "mean_r_generated": 0.21734344370825287,
+     "mean_r_broadcast": 0.29096860411544245,
+     "mean_partial_r": 0.14295055175158178
+    }
+   },
+   "deg_top50": {
+    "types_scored": [
+     "T cell",
+     "B cell",
+     "Monocyte",
+     "NK cell",
+     "DC"
+    ],
+    "types_not_scored": {},
+    "per_type": {
+     "T cell": {
+      "deg_scored": true,
+      "top50_overlap": 1,
+      "top50_jaccard": 0.018867924528301886,
+      "obs_top_list_size": 50,
+      "gen_top_list_size": 4,
+      "top50_overlap_broadcast": 3,
+      "top50_jaccard_broadcast": 0.030927835051546393,
+      "n_cells_obs": 298117,
+      "n_cells_gen": 65670
+     },
+     "B cell": {
+      "deg_scored": true,
+      "top50_overlap": 1,
+      "top50_jaccard": 0.01818181818181818,
+      "obs_top_list_size": 50,
+      "gen_top_list_size": 6,
+      "top50_overlap_broadcast": 2,
+      "top50_jaccard_broadcast": 0.02040816326530612,
+      "n_cells_obs": 99729,
+      "n_cells_gen": 14754
+     },
+     "Monocyte": {
+      "deg_scored": true,
+      "top50_overlap": 3,
+      "top50_jaccard": 0.030927835051546393,
+      "obs_top_list_size": 50,
+      "gen_top_list_size": 50,
+      "top50_overlap_broadcast": 13,
+      "top50_jaccard_broadcast": 0.14942528735632185,
+      "n_cells_obs": 111269,
+      "n_cells_gen": 27348
+     },
+     "NK cell": {
+      "deg_scored": true,
+      "top50_overlap": 0,
+      "top50_jaccard": 0.0,
+      "obs_top_list_size": 50,
+      "gen_top_list_size": 1,
+      "top50_overlap_broadcast": 2,
+      "top50_jaccard_broadcast": 0.02040816326530612,
+      "n_cells_obs": 47832,
+      "n_cells_gen": 6290
+     },
+     "DC": {
+      "deg_scored": true,
+      "top50_overlap": 1,
+      "top50_jaccard": 0.015384615384615385,
+      "obs_top_list_size": 50,
+      "gen_top_list_size": 16,
+      "top50_overlap_broadcast": 1,
+      "top50_jaccard_broadcast": 0.010101010101010102,
+      "n_cells_obs": 3300,
+      "n_cells_gen": 2369
+     }
+    },
+    "sum": 6,
+    "sum_4type": 5,
+    "max": 250,
+    "mean": 1.2,
+    "sum_broadcast": 21,
+    "sum_broadcast_4type": 20,
+    "broadcast_n_significant": 1378,
+    "seconds": 0.12
+   },
+   "seconds": 29.0
+  }
+ },
+ "code_sha256": {
+  "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/tools/score_ft.py": "ec5f40b4cea6d5062d0138f3b934abb6465e34cbfeac869c8a0906924c033aec",
+  "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/paired_finetune_intradataset_20260927/tools/ft_common.py": "4a6366b0f22208cd58714fc0261b8840c643b9d4569704d560cfd93534cbc747",
+  "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/umap/cell_immicro_5arm/tools/stage_b_embed.py": "d056cfdf38fb0e5f478756f370174e993b75d0c973c48e02e21749175a4ca2da",
+  "/home/sj_server_1/PBISC/immun-pbmc-fm/bulk2sc-paper/scripts/evaluate_generated_vs_truth.py": "48b12d794782b840838853e93e0fc55fab00532c4b3698f4588e926b985d2881",
+  "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/utility_studies_final6k_20260925/immicro/tools/immicro_common.py": "c5c30e917148cc5e664e91ea30c11bbdff9b6d1f53a3e0a13e16b1917eff37e9",
+  "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/aifi_scanvi_rebuild_20260916/predict_scanvi.py": "482ee5dcf4d72b54bc36cbb6dcff62e5030bc7afc46cb5fb01eca73a82ee45a5",
+  "/home/sj_server_1/PBISC/immun-pbmc-fm/model_8_sparse_data_diffusion/plans/aifi_scanvi_rebuild_20260916/poisson_readout.py": "8bfa080942fbcdc89ae23b276524a1a83266fb8cbc56b9af279e50f82e6cde38"
+ },
+ "seconds": 38.3
+}