From a4a9e4e2de7d9c2525f9b17480bb60240e713648 Mon Sep 17 00:00:00 2001 From: Jammy2211 Date: Sun, 13 Sep 2026 12:04:53 +0100 Subject: [PATCH 1/4] =?UTF-8?q?cortex:=20euclid=5Fsersics=20birth=20?= =?UTF-8?q?=E2=80=94=20row,=20ledger=20and=20the=20core=20run=20342696?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Registers the science project that asks why the lens-light Sersic index piles up at the prior edge n = 5 in the June Euclid catalogue (61 % of 2983 fits above 4.5; 67 of the chosen 100). The row carries the site copy of euclid_strong_lens_modeling_pipeline at /mnt/c/Users/Jammy/Science/euclid_sersics, RAL /mnt/ral/jnightin/euclid_sersics and its own hpc/sync; partition ral. The ledger opens with the birth note of 2026-09-11 and the core two-stage run 342696_[0-99] (vis_lp + vis_pix over the 100-lens dr1_sep1_sersics sample), submitted 09-11 and finished 100/100 with task walls 2:47 to 20:58, its outputs pulled on 09-12. No result and no lesson: the human has not ruled. Board re-rendered in the same commit. Co-Authored-By: Claude Fable 5.1 Claude-Session: https://claude.ai/code/session_01EehLEWoRRsmnLys4aHj5W4 --- dashboard.html | 20 +++++++++++++++++++- dashboard.md | 34 +++++++++++++++++++++++++++++++++- projects.yaml | 14 ++++++++++++++ projects/euclid_sersics.md | 20 ++++++++++++++++++++ 4 files changed, 86 insertions(+), 2 deletions(-) create mode 100644 projects/euclid_sersics.md diff --git a/dashboard.html b/dashboard.html index 0d1ab40..29e22f3 100644 --- a/dashboard.html +++ b/dashboard.html @@ -227,7 +227,7 @@

PyAutoCortexDashboard

Question. Run. Rule.

One ledger per science project: what is on the cluster right now, what the human said they would do next, and the last few things they wrote down. Nothing here is inferred from a result β€” the board hands you the check-in and the resume prompt; what a run meant is yours to log.

- +

Last updated 2026-09-13.

markdown version Β· GitHub Page

check in since last time

@@ -237,6 +237,7 @@ ProjectRunningOpenLast update subhalo_validation052026-09-07 euclid_dr1_prelim132026-09-12 +euclid_sersics002026-09-12 analytic_gaussian002026-09-10 ep_toy_gaussian022026-09-10 slope_hierarchy_scale042026-09-09 @@ -292,6 +293,23 @@

euclid_dr1_prelim β€” Refit the ten DR1-prelim tiles and rebuild the catalog

resume euclid_dr1_prelim

+

euclid_sersics β€” Why does the lens-light Sersic index pile up at the n=5 prior edge

+

active Β· ral Β· euclid_strong_lens_modeling_pipeline#74 Β· projects/euclid_sersics.md Β· local /mnt/c/Users/Jammy/Science/euclid_sersics Β· RAL /mnt/ral/jnightin/euclid_sersics

+

Now

+

342696 fitted all 100 dr1_sep1_sersics lenses through vis_lp and vis_pix (100/100 tasks, walls 2:47 to 20:58) and the outputs are pulled to the local clone; that run is what the four Sersic-stage variants reuse rather than re-derive.
The variants array β€” baseline, wide_n, central_noise and sersic_point, the four run sequentially inside each of 100 array tasks β€” is being submitted to the cluster now (4 cpu / 14 GB / 18 h, partition ral); its pipeline code is euclid_strong_lens_modeling_pipeline#75 (open, CI green, awaiting the human's merge) and the Mind task is sersic-variants.
Next: pull, scrape each variant with catalogue/scripts/lens_sersic.py --unique_tag sersic_lens_model_<variant>, run the per-variant analysis script, and the human rules on the four n distributions.

+

Runs

+

nothing on the cluster

+

Last 5

+
    +
  • 2026-09-12 β€” run β€” 342696_[0-99] finished β€” wall 20:58: euclid_sersics_two_stage: the core two-stage fit (vis_lp then vis_pix) of the 100-lens dr1_sep1_sersics sample, array 0-99, one lens per task, 8 cpu / 64 GB / 36 h on partition ral, submitted 11:43 BST;…
    • +
  • 2026-09-11 β€” run β€” 342696_[0-99] submitted: euclid_sersics_two_stage: the core two-stage fit (vis_lp then vis_pix) of the 100-lens dr1_sep1_sersics sample, array 0-99, one lens per task, 8 cpu / 64 GB / 36 h on partition ral, submitted 11:43 BST; it is the r…
    • +
  • 2026-09-11 β€” note β€” project born ~11:45 BST once the human confirmed the euclid_dr1 sep1 results good: 100 unanimous-success lenses from the euclid inspection tables, remodelled on the sep1 Segmentation.zip data, to explain why the lens-light Sersic index pil…
    • +
  • 2026-09-11 β€” note β€” ledger opened
    • +
+

full log

+

resume euclid_sersics

+
+

analytic_gaussian β€” Graphical and EP against a closed-form Gaussian posterior, at ensemble scale

active Β· ral Β· PyAutoCortex#34 Β· projects/analytic_gaussian.md Β· local /mnt/c/Users/Jammy/Science/analytic_gaussian Β· RAL /mnt/ral/jnightin/analytic_gaussian

Now

diff --git a/dashboard.md b/dashboard.md index 6aabf41..353ced1 100644 --- a/dashboard.md +++ b/dashboard.md @@ -12,7 +12,7 @@ One ledger per science project: what is on the cluster right now, what the human |-------|------:| | [Running](#projects) | 1 | | [Open](#projects) | 20 | -| [Projects](#projects) | 7 | +| [Projects](#projects) | 8 |
πŸ“‹ check in since last time @@ -30,6 +30,7 @@ One ledger per science project: what is on the cluster right now, what the human |---|---:|---:|---| | subhalo_validation | 0 | 5 | 2026-09-07 | | euclid_dr1_prelim | 1 | 3 | 2026-09-12 | +| euclid_sersics | 0 | 0 | 2026-09-12 | | analytic_gaussian | 0 | 0 | 2026-09-10 | | ep_toy_gaussian | 0 | 2 | 2026-09-10 | | slope_hierarchy_scale | 0 | 4 | 2026-09-09 | @@ -107,6 +108,37 @@ Next: pull, bundle with output_sed, compare magnitudes (Β΅Jy fluxes) against the
+### euclid_sersics β€” Why does the lens-light Sersic index pile up at the n=5 prior edge + +active Β· ral Β· [euclid_strong_lens_modeling_pipeline#74](https://github.com/PyAutoLabs/euclid_strong_lens_modeling_pipeline/issues/74) Β· [projects/euclid_sersics.md](https://github.com/PyAutoLabs/PyAutoCortex/blob/main/projects/euclid_sersics.md) Β· local `/mnt/c/Users/Jammy/Science/euclid_sersics` Β· RAL `/mnt/ral/jnightin/euclid_sersics` + +**Now** + +342696 fitted all 100 dr1_sep1_sersics lenses through vis_lp and vis_pix (100/100 tasks, walls 2:47 to 20:58) and the outputs are pulled to the local clone; that run is what the four Sersic-stage variants reuse rather than re-derive. +The variants array β€” baseline, wide_n, central_noise and sersic_point, the four run sequentially inside each of 100 array tasks β€” is being submitted to the cluster now (4 cpu / 14 GB / 18 h, partition ral); its pipeline code is euclid_strong_lens_modeling_pipeline#75 (open, CI green, awaiting the human's merge) and the Mind task is sersic-variants. +Next: pull, scrape each variant with catalogue/scripts/lens_sersic.py --unique_tag sersic_lens_model_, run the per-variant analysis script, and the human rules on the four n distributions. + +**Runs** + +- _nothing on the cluster_ + +**Last 5** + +- 2026-09-12 β€” *run* β€” 342696_[0-99] finished β€” wall 20:58: euclid_sersics_two_stage: the core two-stage fit (vis_lp then vis_pix) of the 100-lens dr1_sep1_sersics sample, array 0-99, one lens per task, 8 cpu / 64 GB / 36 h on partition ral, submitted 11:43 BST;… +- 2026-09-11 β€” *run* β€” 342696_[0-99] submitted: euclid_sersics_two_stage: the core two-stage fit (vis_lp then vis_pix) of the 100-lens dr1_sep1_sersics sample, array 0-99, one lens per task, 8 cpu / 64 GB / 36 h on partition ral, submitted 11:43 BST; it is the r… +- 2026-09-11 β€” *note* β€” project born ~11:45 BST once the human confirmed the euclid_dr1 sep1 results good: 100 unanimous-success lenses from the euclid inspection tables, remodelled on the sep1 Segmentation.zip data, to explain why the lens-light Sersic index pil… +- 2026-09-11 β€” *note* β€” ledger opened + +[full log](https://github.com/PyAutoLabs/PyAutoCortex/blob/main/projects/euclid_sersics.md) + +
πŸ“‹ resume euclid_sersics + +``` +/cortex β€” resume euclid_sersics: read PyAutoCortex projects/euclid_sersics.md (Now, Runs, Log) and then /mnt/c/Users/Jammy/Science/euclid_sersics/wiki/project/state.md and the assistant autolens_assistant's AGENTS.md; tell me where I left off and what I said I would do next. Submit nothing and log nothing until I say. +``` + +
+ ### analytic_gaussian β€” Graphical and EP against a closed-form Gaussian posterior, at ensemble scale active Β· ral Β· [PyAutoCortex#34](https://github.com/PyAutoLabs/PyAutoCortex/issues/34) Β· [projects/analytic_gaussian.md](https://github.com/PyAutoLabs/PyAutoCortex/blob/main/projects/analytic_gaussian.md) Β· local `/mnt/c/Users/Jammy/Science/analytic_gaussian` Β· RAL `/mnt/ral/jnightin/analytic_gaussian` diff --git a/projects.yaml b/projects.yaml index d81b859..e8fc5a4 100644 --- a/projects.yaml +++ b/projects.yaml @@ -101,6 +101,20 @@ euclid_dr1_prelim: status: active note: "created by euclid-dr1-prep phase 4 (a Cortex task after migration); every path is the planned one; active 2026-09-07: phase 4 launched (jobs 342301_[0-9])" +euclid_sersics: + remote: none + local_path: /mnt/c/Users/Jammy/Science/euclid_sersics + ral_root: /mnt/ral/jnightin/euclid_sersics + mirror: none + sync_cli: hpc/sync + sync_verbs: [push, pull, logs, sync, push-data-init, pull-full, status, submit, push-submit, jobs, sacct, cancel, wait-and-pull, tail, du, check, clear-logs] + ledger: wiki/project/state.md + assistant: autolens_assistant + witness_file: results/**/*.json + partition: ral + status: active + note: "born 2026-09-11: why the lens-light Sersic index piles up at the prior edge n = 5 in the June Euclid catalogue (61 % of 2983 fits above 4.5; 67 of the chosen 100); site copy of PyAutoLabs/euclid_strong_lens_modeling_pipeline, sample dataset/dr1_sep1_sersics (100 unanimous-success lenses on the sep1 Segmentation.zip data, join table sample/lens_map.csv); remote none until the human creates one" + analytic_gaussian: remote: none local_path: /mnt/c/Users/Jammy/Science/analytic_gaussian diff --git a/projects/euclid_sersics.md b/projects/euclid_sersics.md new file mode 100644 index 0000000..db50e70 --- /dev/null +++ b/projects/euclid_sersics.md @@ -0,0 +1,20 @@ +# euclid_sersics β€” Why does the lens-light Sersic index pile up at the n=5 prior edge + +Project: euclid_sersics +Issue: euclid_strong_lens_modeling_pipeline#74 + +## Now + +342696 fitted all 100 dr1_sep1_sersics lenses through vis_lp and vis_pix (100/100 tasks, walls 2:47 to 20:58) and the outputs are pulled to the local clone; that run is what the four Sersic-stage variants reuse rather than re-derive. +The variants array β€” baseline, wide_n, central_noise and sersic_point, the four run sequentially inside each of 100 array tasks β€” is being submitted to the cluster now (4 cpu / 14 GB / 18 h, partition ral); its pipeline code is euclid_strong_lens_modeling_pipeline#75 (open, CI green, awaiting the human's merge) and the Mind task is sersic-variants. +Next: pull, scrape each variant with catalogue/scripts/lens_sersic.py --unique_tag sersic_lens_model_, run the per-variant analysis script, and the human rules on the four n distributions. + +## Runs + + +## Log + +- 2026-09-12 β€” run β€” 342696_[0-99] finished β€” wall 20:58: euclid_sersics_two_stage: the core two-stage fit (vis_lp then vis_pix) of the 100-lens dr1_sep1_sersics sample, array 0-99, one lens per task, 8 cpu / 64 GB / 36 h on partition ral, submitted 11:43 BST; it is the run the four Sersic-stage variants reuse rather than re-derive β€” 100/100 tasks completed both stages (vis_lp and vis_pix); task walls 2:47 to 20:58; outputs pulled to the local clone on 2026-09-12 +- 2026-09-11 β€” run β€” 342696_[0-99] submitted: euclid_sersics_two_stage: the core two-stage fit (vis_lp then vis_pix) of the 100-lens dr1_sep1_sersics sample, array 0-99, one lens per task, 8 cpu / 64 GB / 36 h on partition ral, submitted 11:43 BST; it is the run the four Sersic-stage variants reuse rather than re-derive +- 2026-09-11 β€” note β€” project born ~11:45 BST once the human confirmed the euclid_dr1 sep1 results good: 100 unanimous-success lenses from the euclid inspection tables, remodelled on the sep1 Segmentation.zip data, to explain why the lens-light Sersic index piles up at the prior edge n = 5 (June catalogue: 61 % of 2983 fits above 4.5; 67 of the chosen 100 above 4.5). Local clone /mnt/c/Users/Jammy/Science/euclid_sersics (a site copy of PyAutoLabs/euclid_strong_lens_modeling_pipeline, its own journal at wiki/project/state.md), RAL root /mnt/ral/jnightin/euclid_sersics, sync CLI hpc/sync, sample dataset/dr1_sep1_sersics (100 tiles) with the June-to-sep1 join in sample/lens_map.csv. The plan is the core two-stage run first, then the Sersic stage rerun in four variants β€” baseline (the June recipe on the new data), wide_n (lens n prior Uniform(0.5, 10)), central_noise (multiplicative Gaussian noise inflation A = 9, sigma = 0.17 arcsec at the vis_lp lens-light centre) and sersic_point (Sersic + 5-Gaussian point MGE, free centre, +4 params) β€” with the human ruling on the per-variant n distributions afterwards; nothing is pre-registered as a verdict. +- 2026-09-11 β€” note β€” ledger opened From b78add7f0a68dd2d664579a02585f77f3484040c Mon Sep 17 00:00:00 2001 From: Jammy2211 Date: Sun, 13 Sep 2026 12:11:00 +0100 Subject: [PATCH 2/4] cortex: euclid_sersics variants array 342808 submitted MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The Sersic stage rerun in four variants β€” baseline, wide_n, central_noise and sersic_point β€” as job 342808_[0-99], submitted 2026-09-13 ~12:10 BST: one lens per array task with the four variants run sequentially inside it, 4 cpu / 14 GB / 18 h on partition ral, writing sersic_lens_model_{baseline,wide_n, central_noise,sersic_point}. All four reuse 342696's vis_lp result rather than re-derive it. The row's witness_file becomes catalogue/**/*.csv (the same glob as the euclid row): this project's witness is the per-variant `catalogue/scripts/lens_sersic.py --unique_tag` scrape, not a results JSON. Board re-rendered in the same commit. Co-Authored-By: Claude Fable 5.1 Claude-Session: https://claude.ai/code/session_01EehLEWoRRsmnLys4aHj5W4 --- dashboard.html | 11 +++++++---- dashboard.md | 9 +++++---- projects.yaml | 2 +- projects/euclid_sersics.md | 4 +++- 4 files changed, 16 insertions(+), 10 deletions(-) diff --git a/dashboard.html b/dashboard.html index 29e22f3..e4beb73 100644 --- a/dashboard.html +++ b/dashboard.html @@ -227,7 +227,7 @@

PyAutoCortexDashboard

Question. Run. Rule.

One ledger per science project: what is on the cluster right now, what the human said they would do next, and the last few things they wrote down. Nothing here is inferred from a result β€” the board hands you the check-in and the resume prompt; what a run meant is yours to log.

-
  • 1Running
  • 20Open
  • 8Projects
+
  • 1Running
  • 21Open
  • 8Projects

Last updated 2026-09-13.

markdown version Β· GitHub Page

check in since last time

@@ -237,7 +237,7 @@ ProjectRunningOpenLast update subhalo_validation052026-09-07 euclid_dr1_prelim132026-09-12 -euclid_sersics002026-09-12 +euclid_sersics012026-09-13 analytic_gaussian002026-09-10 ep_toy_gaussian022026-09-10 slope_hierarchy_scale042026-09-09 @@ -296,11 +296,14 @@

euclid_dr1_prelim β€” Refit the ten DR1-prelim tiles and rebuild the catalog

euclid_sersics β€” Why does the lens-light Sersic index pile up at the n=5 prior edge

active Β· ral Β· euclid_strong_lens_modeling_pipeline#74 Β· projects/euclid_sersics.md Β· local /mnt/c/Users/Jammy/Science/euclid_sersics Β· RAL /mnt/ral/jnightin/euclid_sersics

Now

-

342696 fitted all 100 dr1_sep1_sersics lenses through vis_lp and vis_pix (100/100 tasks, walls 2:47 to 20:58) and the outputs are pulled to the local clone; that run is what the four Sersic-stage variants reuse rather than re-derive.
The variants array β€” baseline, wide_n, central_noise and sersic_point, the four run sequentially inside each of 100 array tasks β€” is being submitted to the cluster now (4 cpu / 14 GB / 18 h, partition ral); its pipeline code is euclid_strong_lens_modeling_pipeline#75 (open, CI green, awaiting the human's merge) and the Mind task is sersic-variants.
Next: pull, scrape each variant with catalogue/scripts/lens_sersic.py --unique_tag sersic_lens_model_<variant>, run the per-variant analysis script, and the human rules on the four n distributions.

+

342696 fitted all 100 dr1_sep1_sersics lenses through vis_lp and vis_pix (100/100 tasks, walls 2:47 to 20:58) and the outputs are pulled to the local clone; that run is what the four Sersic-stage variants reuse rather than re-derive.
342808_[0-99] is on the cluster: the variants array β€” baseline, wide_n, central_noise and sersic_point, the four run sequentially inside each of 100 array tasks β€” submitted 2026-09-13 ~12:10 BST (4 cpu / 14 GB / 18 h, partition ral); its pipeline code is euclid_strong_lens_modeling_pipeline#75 (open, CI green, awaiting the human's merge) and the Mind task is sersic-variants.
Next: pull, scrape each variant with catalogue/scripts/lens_sersic.py --unique_tag sersic_lens_model_<variant>, run the per-variant analysis script, and the human rules on the four n distributions.

Runs

-

nothing on the cluster

+
    +
  • 342808_[0-99] β€” open β€” ral β€” 2026-09-13 β€” euclid_sersics_variants: the Sersic stage rerun in four variants β€” baseline, wide_n, central_noise, sersic_point β€” run sequentially inside each of 100 array ta…
    • +

Last 5

    +
  • 2026-09-13 β€” run β€” 342808_[0-99] submitted: euclid_sersics_variants: the Sersic stage rerun in four variants β€” baseline, wide_n, central_noise, sersic_point β€” run sequentially inside each of 100 array tasks (one lens per task), array 0-99, 4 cpu / 14 GB / 18…
  • 2026-09-12 β€” run β€” 342696_[0-99] finished β€” wall 20:58: euclid_sersics_two_stage: the core two-stage fit (vis_lp then vis_pix) of the 100-lens dr1_sep1_sersics sample, array 0-99, one lens per task, 8 cpu / 64 GB / 36 h on partition ral, submitted 11:43 BST;…
  • 2026-09-11 β€” run β€” 342696_[0-99] submitted: euclid_sersics_two_stage: the core two-stage fit (vis_lp then vis_pix) of the 100-lens dr1_sep1_sersics sample, array 0-99, one lens per task, 8 cpu / 64 GB / 36 h on partition ral, submitted 11:43 BST; it is the r…
  • 2026-09-11 β€” note β€” project born ~11:45 BST once the human confirmed the euclid_dr1 sep1 results good: 100 unanimous-success lenses from the euclid inspection tables, remodelled on the sep1 Segmentation.zip data, to explain why the lens-light Sersic index pil…
    • diff --git a/dashboard.md b/dashboard.md index 353ced1..3b2d43b 100644 --- a/dashboard.md +++ b/dashboard.md @@ -11,7 +11,7 @@ One ledger per science project: what is on the cluster right now, what the human | Where | Count | |-------|------:| | [Running](#projects) | 1 | -| [Open](#projects) | 20 | +| [Open](#projects) | 21 | | [Projects](#projects) | 8 |
    πŸ“‹ check in since last time @@ -30,7 +30,7 @@ One ledger per science project: what is on the cluster right now, what the human |---|---:|---:|---| | subhalo_validation | 0 | 5 | 2026-09-07 | | euclid_dr1_prelim | 1 | 3 | 2026-09-12 | -| euclid_sersics | 0 | 0 | 2026-09-12 | +| euclid_sersics | 0 | 1 | 2026-09-13 | | analytic_gaussian | 0 | 0 | 2026-09-10 | | ep_toy_gaussian | 0 | 2 | 2026-09-10 | | slope_hierarchy_scale | 0 | 4 | 2026-09-09 | @@ -115,15 +115,16 @@ active Β· ral Β· [euclid_strong_lens_modeling_pipeline#74](https://github.com/Py **Now** 342696 fitted all 100 dr1_sep1_sersics lenses through vis_lp and vis_pix (100/100 tasks, walls 2:47 to 20:58) and the outputs are pulled to the local clone; that run is what the four Sersic-stage variants reuse rather than re-derive. -The variants array β€” baseline, wide_n, central_noise and sersic_point, the four run sequentially inside each of 100 array tasks β€” is being submitted to the cluster now (4 cpu / 14 GB / 18 h, partition ral); its pipeline code is euclid_strong_lens_modeling_pipeline#75 (open, CI green, awaiting the human's merge) and the Mind task is sersic-variants. +342808_[0-99] is on the cluster: the variants array β€” baseline, wide_n, central_noise and sersic_point, the four run sequentially inside each of 100 array tasks β€” submitted 2026-09-13 ~12:10 BST (4 cpu / 14 GB / 18 h, partition ral); its pipeline code is euclid_strong_lens_modeling_pipeline#75 (open, CI green, awaiting the human's merge) and the Mind task is sersic-variants. Next: pull, scrape each variant with catalogue/scripts/lens_sersic.py --unique_tag sersic_lens_model_, run the per-variant analysis script, and the human rules on the four n distributions. **Runs** -- _nothing on the cluster_ +- 342808_[0-99] β€” open β€” ral β€” 2026-09-13 β€” euclid_sersics_variants: the Sersic stage rerun in four variants β€” baseline, wide_n, central_noise, sersic_point β€” run sequentially inside each of 100 array ta… **Last 5** +- 2026-09-13 β€” *run* β€” 342808_[0-99] submitted: euclid_sersics_variants: the Sersic stage rerun in four variants β€” baseline, wide_n, central_noise, sersic_point β€” run sequentially inside each of 100 array tasks (one lens per task), array 0-99, 4 cpu / 14 GB / 18… - 2026-09-12 β€” *run* β€” 342696_[0-99] finished β€” wall 20:58: euclid_sersics_two_stage: the core two-stage fit (vis_lp then vis_pix) of the 100-lens dr1_sep1_sersics sample, array 0-99, one lens per task, 8 cpu / 64 GB / 36 h on partition ral, submitted 11:43 BST;… - 2026-09-11 β€” *run* β€” 342696_[0-99] submitted: euclid_sersics_two_stage: the core two-stage fit (vis_lp then vis_pix) of the 100-lens dr1_sep1_sersics sample, array 0-99, one lens per task, 8 cpu / 64 GB / 36 h on partition ral, submitted 11:43 BST; it is the r… - 2026-09-11 β€” *note* β€” project born ~11:45 BST once the human confirmed the euclid_dr1 sep1 results good: 100 unanimous-success lenses from the euclid inspection tables, remodelled on the sep1 Segmentation.zip data, to explain why the lens-light Sersic index pil… diff --git a/projects.yaml b/projects.yaml index e8fc5a4..f5d7da3 100644 --- a/projects.yaml +++ b/projects.yaml @@ -110,7 +110,7 @@ euclid_sersics: sync_verbs: [push, pull, logs, sync, push-data-init, pull-full, status, submit, push-submit, jobs, sacct, cancel, wait-and-pull, tail, du, check, clear-logs] ledger: wiki/project/state.md assistant: autolens_assistant - witness_file: results/**/*.json + witness_file: catalogue/**/*.csv partition: ral status: active note: "born 2026-09-11: why the lens-light Sersic index piles up at the prior edge n = 5 in the June Euclid catalogue (61 % of 2983 fits above 4.5; 67 of the chosen 100); site copy of PyAutoLabs/euclid_strong_lens_modeling_pipeline, sample dataset/dr1_sep1_sersics (100 unanimous-success lenses on the sep1 Segmentation.zip data, join table sample/lens_map.csv); remote none until the human creates one" diff --git a/projects/euclid_sersics.md b/projects/euclid_sersics.md index db50e70..6e38ee8 100644 --- a/projects/euclid_sersics.md +++ b/projects/euclid_sersics.md @@ -6,14 +6,16 @@ Issue: euclid_strong_lens_modeling_pipeline#74 ## Now 342696 fitted all 100 dr1_sep1_sersics lenses through vis_lp and vis_pix (100/100 tasks, walls 2:47 to 20:58) and the outputs are pulled to the local clone; that run is what the four Sersic-stage variants reuse rather than re-derive. -The variants array β€” baseline, wide_n, central_noise and sersic_point, the four run sequentially inside each of 100 array tasks β€” is being submitted to the cluster now (4 cpu / 14 GB / 18 h, partition ral); its pipeline code is euclid_strong_lens_modeling_pipeline#75 (open, CI green, awaiting the human's merge) and the Mind task is sersic-variants. +342808_[0-99] is on the cluster: the variants array β€” baseline, wide_n, central_noise and sersic_point, the four run sequentially inside each of 100 array tasks β€” submitted 2026-09-13 ~12:10 BST (4 cpu / 14 GB / 18 h, partition ral); its pipeline code is euclid_strong_lens_modeling_pipeline#75 (open, CI green, awaiting the human's merge) and the Mind task is sersic-variants. Next: pull, scrape each variant with catalogue/scripts/lens_sersic.py --unique_tag sersic_lens_model_, run the per-variant analysis script, and the human rules on the four n distributions. ## Runs +- 342808_[0-99] β€” open β€” ral β€” 2026-09-13 β€” euclid_sersics_variants: the Sersic stage rerun in four variants β€” baseline, wide_n, central_noise, sersic_point β€” run sequentially inside each of 100 array tasks (one lens per task), array 0-99, 4 cpu / 14 GB / 18 h on partition ral, submitted ~12:10 BST; output tags sersic_lens_model_{baseline,wide_n,central_noise,sersic_point}; all four reuse 342696's vis_lp result; pipeline code euclid_strong_lens_modeling_pipeline#75 (open, CI green) ## Log +- 2026-09-13 β€” run β€” 342808_[0-99] submitted: euclid_sersics_variants: the Sersic stage rerun in four variants β€” baseline, wide_n, central_noise, sersic_point β€” run sequentially inside each of 100 array tasks (one lens per task), array 0-99, 4 cpu / 14 GB / 18 h on partition ral, submitted ~12:10 BST; output tags sersic_lens_model_{baseline,wide_n,central_noise,sersic_point}; all four reuse 342696's vis_lp result; pipeline code euclid_strong_lens_modeling_pipeline#75 (open, CI green) - 2026-09-12 β€” run β€” 342696_[0-99] finished β€” wall 20:58: euclid_sersics_two_stage: the core two-stage fit (vis_lp then vis_pix) of the 100-lens dr1_sep1_sersics sample, array 0-99, one lens per task, 8 cpu / 64 GB / 36 h on partition ral, submitted 11:43 BST; it is the run the four Sersic-stage variants reuse rather than re-derive β€” 100/100 tasks completed both stages (vis_lp and vis_pix); task walls 2:47 to 20:58; outputs pulled to the local clone on 2026-09-12 - 2026-09-11 β€” run β€” 342696_[0-99] submitted: euclid_sersics_two_stage: the core two-stage fit (vis_lp then vis_pix) of the 100-lens dr1_sep1_sersics sample, array 0-99, one lens per task, 8 cpu / 64 GB / 36 h on partition ral, submitted 11:43 BST; it is the run the four Sersic-stage variants reuse rather than re-derive - 2026-09-11 β€” note β€” project born ~11:45 BST once the human confirmed the euclid_dr1 sep1 results good: 100 unanimous-success lenses from the euclid inspection tables, remodelled on the sep1 Segmentation.zip data, to explain why the lens-light Sersic index piles up at the prior edge n = 5 (June catalogue: 61 % of 2983 fits above 4.5; 67 of the chosen 100 above 4.5). Local clone /mnt/c/Users/Jammy/Science/euclid_sersics (a site copy of PyAutoLabs/euclid_strong_lens_modeling_pipeline, its own journal at wiki/project/state.md), RAL root /mnt/ral/jnightin/euclid_sersics, sync CLI hpc/sync, sample dataset/dr1_sep1_sersics (100 tiles) with the June-to-sep1 join in sample/lens_map.csv. The plan is the core two-stage run first, then the Sersic stage rerun in four variants β€” baseline (the June recipe on the new data), wide_n (lens n prior Uniform(0.5, 10)), central_noise (multiplicative Gaussian noise inflation A = 9, sigma = 0.17 arcsec at the vis_lp lens-light centre) and sersic_point (Sersic + 5-Gaussian point MGE, free centre, +4 params) β€” with the human ruling on the per-variant n distributions afterwards; nothing is pre-registered as a verdict. From f796a0f64c8dce8cb67e2697478f37734f21e7c9 Mon Sep 17 00:00:00 2001 From: Jammy2211 Date: Sun, 13 Sep 2026 14:58:03 +0100 Subject: [PATCH 3/4] cortex: euclid_sersics simulated-sample vis_lp 342911 submitted MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit 342911_[0-99] β€” euclid_sersics_sim_vis_lp β€” runs the vis_lp stage (initial_lens_model.py --stage=vis_lp, MGE lens + MGE source, 15 free parameters, identical to the real sample's stage 1) on the new simulated sample dataset/dr1_sep1_sersics_sim: 8 cpu / 64 GB / 36 h per task, partition ral, submitted 2026-09-13 ~14:56 BST. The sample is the 100 real lenses resimulated by scripts/simulator.py --from-result from each lens's own sersic_lens_model_baseline/vis result of 342808, keeping that tile's real PSF stamps, noise and bands. The note records the prior-edge rule taken while the human was out β€” a baseline n within 1 % of the n = 5 bound simulated at n = 3.0, 49 of 100 replaced β€” together with the edge census and the two pipeline defects fixed on the way, so the sample can be re-cut from the per-tile truth.json. 342808 stays open: it has not finished. Board re-rendered in the same commit. Co-Authored-By: Claude Fable 5.1 Claude-Session: https://claude.ai/code/session_01EehLEWoRRsmnLys4aHj5W4 --- dashboard.html | 11 ++++++----- dashboard.md | 13 +++++++------ projects/euclid_sersics.md | 7 +++++-- 3 files changed, 18 insertions(+), 13 deletions(-) diff --git a/dashboard.html b/dashboard.html index e4beb73..f8153a3 100644 --- a/dashboard.html +++ b/dashboard.html @@ -227,7 +227,7 @@

    PyAutoCortexDashboard

    Question. Run. Rule.

    One ledger per science project: what is on the cluster right now, what the human said they would do next, and the last few things they wrote down. Nothing here is inferred from a result β€” the board hands you the check-in and the resume prompt; what a run meant is yours to log.

    -
    • 1Running
    • 21Open
    • 8Projects
    +
    • 1Running
    • 22Open
    • 8Projects

    Last updated 2026-09-13.

    markdown version Β· GitHub Page

    check in since last time

    @@ -237,7 +237,7 @@ ProjectRunningOpenLast update subhalo_validation052026-09-07 euclid_dr1_prelim132026-09-12 -euclid_sersics012026-09-13 +euclid_sersics022026-09-13 analytic_gaussian002026-09-10 ep_toy_gaussian022026-09-10 slope_hierarchy_scale042026-09-09 @@ -296,18 +296,19 @@

    euclid_dr1_prelim β€” Refit the ten DR1-prelim tiles and rebuild the catalog

    euclid_sersics β€” Why does the lens-light Sersic index pile up at the n=5 prior edge

    active Β· ral Β· euclid_strong_lens_modeling_pipeline#74 Β· projects/euclid_sersics.md Β· local /mnt/c/Users/Jammy/Science/euclid_sersics Β· RAL /mnt/ral/jnightin/euclid_sersics

    Now

    -

    342696 fitted all 100 dr1_sep1_sersics lenses through vis_lp and vis_pix (100/100 tasks, walls 2:47 to 20:58) and the outputs are pulled to the local clone; that run is what the four Sersic-stage variants reuse rather than re-derive.
    342808_[0-99] is on the cluster: the variants array β€” baseline, wide_n, central_noise and sersic_point, the four run sequentially inside each of 100 array tasks β€” submitted 2026-09-13 ~12:10 BST (4 cpu / 14 GB / 18 h, partition ral); its pipeline code is euclid_strong_lens_modeling_pipeline#75 (open, CI green, awaiting the human's merge) and the Mind task is sersic-variants.
    Next: pull, scrape each variant with catalogue/scripts/lens_sersic.py --unique_tag sersic_lens_model_<variant>, run the per-variant analysis script, and the human rules on the four n distributions.

    +

    342696 fitted all 100 dr1_sep1_sersics lenses through vis_lp and vis_pix (100/100 tasks, walls 2:47 to 20:58) and the outputs are pulled to the local clone; that run is what the four Sersic-stage variants reuse rather than re-derive.
    Two runs are open on the cluster: 342808_[0-99], the four Sersic-stage variants (baseline, wide_n, central_noise, sersic_point) on the real sample, submitted 09-13 ~12:10 BST and not finished β€” task 72 restarted at 12:10Z and is still on its later variants; and 342911_[0-99], the vis_lp stage on the new simulated sample dataset/dr1_sep1_sersics_sim (the 100 real lenses resimulated from their own 342808 baseline results, truth per tile in <Tile>/truth.json), submitted 09-13 ~14:56 BST.
    Next: pull both; scrape the four real variants with catalogue/scripts/lens_sersic.py --unique_tag sersic_lens_model_<variant> and run the per-variant analysis; then, once 342911 lands, run the same four Sersic variants on the simulated sample so the recovered n can be read against a known truth. The human rules on the n distributions.

    Runs

    • 342808_[0-99] β€” open β€” ral β€” 2026-09-13 β€” euclid_sersics_variants: the Sersic stage rerun in four variants β€” baseline, wide_n, central_noise, sersic_point β€” run sequentially inside each of 100 array ta…
      • +
    • 342911_[0-99] β€” open β€” ral β€” 2026-09-13 β€” euclid_sersics_sim_vis_lp: the vis_lp stage (initial_lens_model.py --stage=vis_lp, MGE lens + MGE source, 15 free parameters, identical to the real sample's st…

    Last 5

      +
    • 2026-09-13 β€” note β€” the simulated sample behind 342911: scripts/simulator.py --from-result rebuilds each of the 100 lenses from its own sersic_lens_model_baseline/vis result of run 342808 (Sersic lens light + Sersic source), keeping that tile's real PSF stamp…
      • +
    • 2026-09-13 β€” run β€” 342911_[0-99] submitted: euclid_sersics_sim_vis_lp: the vis_lp stage (initial_lens_model.py --stage=vis_lp, MGE lens + MGE source, 15 free parameters, identical to the real sample's stage 1) on the new simulated sample dataset/dr1_sep1_ser…
    • 2026-09-13 β€” run β€” 342808_[0-99] submitted: euclid_sersics_variants: the Sersic stage rerun in four variants β€” baseline, wide_n, central_noise, sersic_point β€” run sequentially inside each of 100 array tasks (one lens per task), array 0-99, 4 cpu / 14 GB / 18…
    • 2026-09-12 β€” run β€” 342696_[0-99] finished β€” wall 20:58: euclid_sersics_two_stage: the core two-stage fit (vis_lp then vis_pix) of the 100-lens dr1_sep1_sersics sample, array 0-99, one lens per task, 8 cpu / 64 GB / 36 h on partition ral, submitted 11:43 BST;…
    • 2026-09-11 β€” run β€” 342696_[0-99] submitted: euclid_sersics_two_stage: the core two-stage fit (vis_lp then vis_pix) of the 100-lens dr1_sep1_sersics sample, array 0-99, one lens per task, 8 cpu / 64 GB / 36 h on partition ral, submitted 11:43 BST; it is the r…
      • -
    • 2026-09-11 β€” note β€” project born ~11:45 BST once the human confirmed the euclid_dr1 sep1 results good: 100 unanimous-success lenses from the euclid inspection tables, remodelled on the sep1 Segmentation.zip data, to explain why the lens-light Sersic index pil…
      • -
    • 2026-09-11 β€” note β€” ledger opened

    full log

    resume euclid_sersics

    diff --git a/dashboard.md b/dashboard.md index 3b2d43b..6fcd010 100644 --- a/dashboard.md +++ b/dashboard.md @@ -11,7 +11,7 @@ One ledger per science project: what is on the cluster right now, what the human | Where | Count | |-------|------:| | [Running](#projects) | 1 | -| [Open](#projects) | 21 | +| [Open](#projects) | 22 | | [Projects](#projects) | 8 |
    πŸ“‹ check in since last time @@ -30,7 +30,7 @@ One ledger per science project: what is on the cluster right now, what the human |---|---:|---:|---| | subhalo_validation | 0 | 5 | 2026-09-07 | | euclid_dr1_prelim | 1 | 3 | 2026-09-12 | -| euclid_sersics | 0 | 1 | 2026-09-13 | +| euclid_sersics | 0 | 2 | 2026-09-13 | | analytic_gaussian | 0 | 0 | 2026-09-10 | | ep_toy_gaussian | 0 | 2 | 2026-09-10 | | slope_hierarchy_scale | 0 | 4 | 2026-09-09 | @@ -115,20 +115,21 @@ active Β· ral Β· [euclid_strong_lens_modeling_pipeline#74](https://github.com/Py **Now** 342696 fitted all 100 dr1_sep1_sersics lenses through vis_lp and vis_pix (100/100 tasks, walls 2:47 to 20:58) and the outputs are pulled to the local clone; that run is what the four Sersic-stage variants reuse rather than re-derive. -342808_[0-99] is on the cluster: the variants array β€” baseline, wide_n, central_noise and sersic_point, the four run sequentially inside each of 100 array tasks β€” submitted 2026-09-13 ~12:10 BST (4 cpu / 14 GB / 18 h, partition ral); its pipeline code is euclid_strong_lens_modeling_pipeline#75 (open, CI green, awaiting the human's merge) and the Mind task is sersic-variants. -Next: pull, scrape each variant with catalogue/scripts/lens_sersic.py --unique_tag sersic_lens_model_, run the per-variant analysis script, and the human rules on the four n distributions. +Two runs are open on the cluster: 342808_[0-99], the four Sersic-stage variants (baseline, wide_n, central_noise, sersic_point) on the real sample, submitted 09-13 ~12:10 BST and not finished β€” task 72 restarted at 12:10Z and is still on its later variants; and 342911_[0-99], the vis_lp stage on the new simulated sample dataset/dr1_sep1_sersics_sim (the 100 real lenses resimulated from their own 342808 baseline results, truth per tile in /truth.json), submitted 09-13 ~14:56 BST. +Next: pull both; scrape the four real variants with catalogue/scripts/lens_sersic.py --unique_tag sersic_lens_model_ and run the per-variant analysis; then, once 342911 lands, run the same four Sersic variants on the simulated sample so the recovered n can be read against a known truth. The human rules on the n distributions. **Runs** - 342808_[0-99] β€” open β€” ral β€” 2026-09-13 β€” euclid_sersics_variants: the Sersic stage rerun in four variants β€” baseline, wide_n, central_noise, sersic_point β€” run sequentially inside each of 100 array ta… +- 342911_[0-99] β€” open β€” ral β€” 2026-09-13 β€” euclid_sersics_sim_vis_lp: the vis_lp stage (initial_lens_model.py --stage=vis_lp, MGE lens + MGE source, 15 free parameters, identical to the real sample's st… **Last 5** +- 2026-09-13 β€” *note* β€” the simulated sample behind 342911: scripts/simulator.py --from-result rebuilds each of the 100 lenses from its own sersic_lens_model_baseline/vis result of run 342808 (Sersic lens light + Sersic source), keeping that tile's real PSF stamp… +- 2026-09-13 β€” *run* β€” 342911_[0-99] submitted: euclid_sersics_sim_vis_lp: the vis_lp stage (initial_lens_model.py --stage=vis_lp, MGE lens + MGE source, 15 free parameters, identical to the real sample's stage 1) on the new simulated sample dataset/dr1_sep1_ser… - 2026-09-13 β€” *run* β€” 342808_[0-99] submitted: euclid_sersics_variants: the Sersic stage rerun in four variants β€” baseline, wide_n, central_noise, sersic_point β€” run sequentially inside each of 100 array tasks (one lens per task), array 0-99, 4 cpu / 14 GB / 18… - 2026-09-12 β€” *run* β€” 342696_[0-99] finished β€” wall 20:58: euclid_sersics_two_stage: the core two-stage fit (vis_lp then vis_pix) of the 100-lens dr1_sep1_sersics sample, array 0-99, one lens per task, 8 cpu / 64 GB / 36 h on partition ral, submitted 11:43 BST;… - 2026-09-11 β€” *run* β€” 342696_[0-99] submitted: euclid_sersics_two_stage: the core two-stage fit (vis_lp then vis_pix) of the 100-lens dr1_sep1_sersics sample, array 0-99, one lens per task, 8 cpu / 64 GB / 36 h on partition ral, submitted 11:43 BST; it is the r… -- 2026-09-11 β€” *note* β€” project born ~11:45 BST once the human confirmed the euclid_dr1 sep1 results good: 100 unanimous-success lenses from the euclid inspection tables, remodelled on the sep1 Segmentation.zip data, to explain why the lens-light Sersic index pil… -- 2026-09-11 β€” *note* β€” ledger opened [full log](https://github.com/PyAutoLabs/PyAutoCortex/blob/main/projects/euclid_sersics.md) diff --git a/projects/euclid_sersics.md b/projects/euclid_sersics.md index 6e38ee8..cc472a8 100644 --- a/projects/euclid_sersics.md +++ b/projects/euclid_sersics.md @@ -6,15 +6,18 @@ Issue: euclid_strong_lens_modeling_pipeline#74 ## Now 342696 fitted all 100 dr1_sep1_sersics lenses through vis_lp and vis_pix (100/100 tasks, walls 2:47 to 20:58) and the outputs are pulled to the local clone; that run is what the four Sersic-stage variants reuse rather than re-derive. -342808_[0-99] is on the cluster: the variants array β€” baseline, wide_n, central_noise and sersic_point, the four run sequentially inside each of 100 array tasks β€” submitted 2026-09-13 ~12:10 BST (4 cpu / 14 GB / 18 h, partition ral); its pipeline code is euclid_strong_lens_modeling_pipeline#75 (open, CI green, awaiting the human's merge) and the Mind task is sersic-variants. -Next: pull, scrape each variant with catalogue/scripts/lens_sersic.py --unique_tag sersic_lens_model_, run the per-variant analysis script, and the human rules on the four n distributions. +Two runs are open on the cluster: 342808_[0-99], the four Sersic-stage variants (baseline, wide_n, central_noise, sersic_point) on the real sample, submitted 09-13 ~12:10 BST and not finished β€” task 72 restarted at 12:10Z and is still on its later variants; and 342911_[0-99], the vis_lp stage on the new simulated sample dataset/dr1_sep1_sersics_sim (the 100 real lenses resimulated from their own 342808 baseline results, truth per tile in /truth.json), submitted 09-13 ~14:56 BST. +Next: pull both; scrape the four real variants with catalogue/scripts/lens_sersic.py --unique_tag sersic_lens_model_ and run the per-variant analysis; then, once 342911 lands, run the same four Sersic variants on the simulated sample so the recovered n can be read against a known truth. The human rules on the n distributions. ## Runs - 342808_[0-99] β€” open β€” ral β€” 2026-09-13 β€” euclid_sersics_variants: the Sersic stage rerun in four variants β€” baseline, wide_n, central_noise, sersic_point β€” run sequentially inside each of 100 array tasks (one lens per task), array 0-99, 4 cpu / 14 GB / 18 h on partition ral, submitted ~12:10 BST; output tags sersic_lens_model_{baseline,wide_n,central_noise,sersic_point}; all four reuse 342696's vis_lp result; pipeline code euclid_strong_lens_modeling_pipeline#75 (open, CI green) +- 342911_[0-99] β€” open β€” ral β€” 2026-09-13 β€” euclid_sersics_sim_vis_lp: the vis_lp stage (initial_lens_model.py --stage=vis_lp, MGE lens + MGE source, 15 free parameters, identical to the real sample's stage 1) on the new simulated sample dataset/dr1_sep1_sersics_sim, array 0-99, one lens per task, 8 cpu / 64 GB / 36 h on partition ral, submitted ~14:56 BST; the sample is the 100 real lenses resimulated from their own 342808 sersic_lens_model_baseline/vis results, so the recovered n can be read against a known truth ## Log +- 2026-09-13 β€” note β€” the simulated sample behind 342911: scripts/simulator.py --from-result rebuilds each of the 100 lenses from its own sersic_lens_model_baseline/vis result of run 342808 (Sersic lens light + Sersic source), keeping that tile's real PSF stamps, noise and bands, a flat SED and seed = index + 1, into dataset/dr1_sep1_sersics_sim with the per-tile truth at /truth.json and the manifest at sample/sim_manifest.csv. Decision taken while the human was out, so it can be re-cut from truth.json: the prior-edge rule --sersic-index-prior-edge=4.95 --sersic-index-replacement=3.0, i.e. a baseline n within 1 % of the n = 5 bound is simulated at n = 3.0 instead β€” 49 of the 100 were replaced; the edge census of the inferred n is max 4.999966, 10 at >= 4.999, 35 at >= 4.99, 49 at >= 4.95 and 56 at >= 4.9. Two defects were fixed on the way, on pipeline branch feature/simulator-from-result-linear (tip 0f4aed9, issue euclid_strong_lens_modeling_pipeline#77, PR parked on Heart RED, merged into the science clone): --from-result rebuilt a dark tracer from linear light profiles and now reads files/tracer.json, and PyAutoArray's simulator draws Poisson noise unconditionally so FFT round-off raised on 4 tiles, clipped pipeline-side with the library bug filed as a Mind draft prompt. +- 2026-09-13 β€” run β€” 342911_[0-99] submitted: euclid_sersics_sim_vis_lp: the vis_lp stage (initial_lens_model.py --stage=vis_lp, MGE lens + MGE source, 15 free parameters, identical to the real sample's stage 1) on the new simulated sample dataset/dr1_sep1_sersics_sim, array 0-99, one lens per task, 8 cpu / 64 GB / 36 h on partition ral, submitted ~14:56 BST; the sample is the 100 real lenses resimulated from their own 342808 sersic_lens_model_baseline/vis results, so the recovered n can be read against a known truth - 2026-09-13 β€” run β€” 342808_[0-99] submitted: euclid_sersics_variants: the Sersic stage rerun in four variants β€” baseline, wide_n, central_noise, sersic_point β€” run sequentially inside each of 100 array tasks (one lens per task), array 0-99, 4 cpu / 14 GB / 18 h on partition ral, submitted ~12:10 BST; output tags sersic_lens_model_{baseline,wide_n,central_noise,sersic_point}; all four reuse 342696's vis_lp result; pipeline code euclid_strong_lens_modeling_pipeline#75 (open, CI green) - 2026-09-12 β€” run β€” 342696_[0-99] finished β€” wall 20:58: euclid_sersics_two_stage: the core two-stage fit (vis_lp then vis_pix) of the 100-lens dr1_sep1_sersics sample, array 0-99, one lens per task, 8 cpu / 64 GB / 36 h on partition ral, submitted 11:43 BST; it is the run the four Sersic-stage variants reuse rather than re-derive β€” 100/100 tasks completed both stages (vis_lp and vis_pix); task walls 2:47 to 20:58; outputs pulled to the local clone on 2026-09-12 - 2026-09-11 β€” run β€” 342696_[0-99] submitted: euclid_sersics_two_stage: the core two-stage fit (vis_lp then vis_pix) of the 100-lens dr1_sep1_sersics sample, array 0-99, one lens per task, 8 cpu / 64 GB / 36 h on partition ral, submitted 11:43 BST; it is the run the four Sersic-stage variants reuse rather than re-derive From 32ae64b3573d83e2882236e0bf02f7b3b9082fdd Mon Sep 17 00:00:00 2001 From: Jammy2211 Date: Mon, 14 Sep 2026 16:54:48 +0100 Subject: [PATCH 4/4] cortex: euclid_sersics sim-sample Sersic variants 343043 submitted 342808 and 342911 both finished (99/100 and 100/100); 343043 runs the four Sersic variants on the 100 mocks, restoring 342911's vis_lp, so recovered n can be read against truth.json. Co-Authored-By: Claude Opus 5 Claude-Session: https://claude.ai/code/session_01GSDqa2J7vmVKHzyB2RhBKP --- dashboard.html | 19 +++++++++---------- dashboard.md | 22 ++++++++++------------ projects/euclid_sersics.md | 11 ++++++----- 3 files changed, 25 insertions(+), 27 deletions(-) diff --git a/dashboard.html b/dashboard.html index f8153a3..57cb002 100644 --- a/dashboard.html +++ b/dashboard.html @@ -4,7 +4,7 @@ PyAutoCortex Dashboard - +