Publish exact H2 S2 pooled scorer review binding, no postfit launch
This commit is contained in:
1 parent
0436fba83f
commit
8354a2f6c0
22 files changed
+76542
No files matched your search
@@ -1910,5 +1910,212 @@
|
||||
"checkpoint_complete": false,
|
||||
"pooled_score_started": false,
|
||||
"gate_modified": false
|
||||
},
|
||||
"latest_H2_S2_postfit_preparation": {
|
||||
"unix": 1794927271.6056442,
|
||||
"status": "PREPARED_BINDING_DELIVERED_REVIEW_PENDING",
|
||||
"root": "/home/frank14f/MeltMind-runs/phase2-h2-s2-postfit-2026-11-17",
|
||||
"binding_sha256": "cabb17224f3a66c37f2e3089515802105194ad5d1a964e70b4c8ac355ac81e5a",
|
||||
"binding": {
|
||||
"reviewed_commit": "25e79995c7e414a0bb6bc66967e95d60289dc337",
|
||||
"contract_sha256": "0bd03f42c3b2db5bef26708b2ef873edfe674f724b59bc4b4de8641980d8d330",
|
||||
"implementation_proof_sha256": "c53fb8fb7bd134b2da4b6921b3770dc8fa0ade6cd7c67fb606ef981f8c2b93aa",
|
||||
"comparison_contract_sha256": "ea9616faebc17f66c4ce678fbc1358ed2292c423a0b016cc05693ebe52d2ecdf"
|
||||
},
|
||||
"code_files": 63,
|
||||
"synthetic_tests": 23,
|
||||
"review_request": "1794927225.586049",
|
||||
"review_request_link": "https://arcticnimbus.slack.com/archives/C0C78UFKGRM/p1794927225586049",
|
||||
"postfit_started": false,
|
||||
"new_review_present": false,
|
||||
"old_queues": {
|
||||
"phase2-h2-s2-2026-11-17": {
|
||||
"root": "/home/frank14f/MeltMind-runs/phase2-h2-s2-2026-11-17",
|
||||
"live": {
|
||||
"stage": "S2_TRAINING",
|
||||
"unix": 1794927271.289092,
|
||||
"active": [
|
||||
{
|
||||
"fold": 1,
|
||||
"guard_pid": 701953
|
||||
},
|
||||
{
|
||||
"fold": 2,
|
||||
"guard_pid": 701954
|
||||
},
|
||||
{
|
||||
"fold": 3,
|
||||
"guard_pid": 701955
|
||||
},
|
||||
{
|
||||
"fold": 4,
|
||||
"guard_pid": 701956
|
||||
}
|
||||
],
|
||||
"terminal": [],
|
||||
"new_gpu_seconds_charged": 8821.600681224954,
|
||||
"cumulative_h2_gpu_seconds": 27679.815589939943,
|
||||
"prediction_gpu_reserve": 2400,
|
||||
"train_gpu_cap": 122741.78509128501,
|
||||
"epoch_counts": {
|
||||
"1": 20,
|
||||
"2": 20,
|
||||
"3": 25,
|
||||
"4": 20
|
||||
},
|
||||
"rss_bytes": 7660138496,
|
||||
"disk_bytes": 20602880
|
||||
},
|
||||
"complete_present": false,
|
||||
"processes": [
|
||||
{
|
||||
"cmdline": [
|
||||
"/home/frank14f/optane/MeltMind/artifacts/gre24-reproduction-2026-11-15/environment/venv/bin/python",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s2_queue.py",
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-2026-11-17"
|
||||
],
|
||||
"ppid": 2234,
|
||||
"pid": 701941,
|
||||
"status": "sleeping"
|
||||
},
|
||||
{
|
||||
"cmdline": [
|
||||
"/home/frank14f/optane/MeltMind/artifacts/gre24-reproduction-2026-11-15/environment/venv/bin/python",
|
||||
"tools/phase2_h2_s2_guard.py",
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-2026-11-17",
|
||||
"1"
|
||||
],
|
||||
"ppid": 701941,
|
||||
"pid": 701953,
|
||||
"status": "sleeping"
|
||||
},
|
||||
{
|
||||
"cmdline": [
|
||||
"/home/frank14f/optane/MeltMind/artifacts/gre24-reproduction-2026-11-15/environment/venv/bin/python",
|
||||
"tools/phase2_h2_s2_guard.py",
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-2026-11-17",
|
||||
"2"
|
||||
],
|
||||
"ppid": 701941,
|
||||
"pid": 701954,
|
||||
"status": "sleeping"
|
||||
},
|
||||
{
|
||||
"cmdline": [
|
||||
"/home/frank14f/optane/MeltMind/artifacts/gre24-reproduction-2026-11-15/environment/venv/bin/python",
|
||||
"tools/phase2_h2_s2_guard.py",
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-2026-11-17",
|
||||
"3"
|
||||
],
|
||||
"ppid": 701941,
|
||||
"pid": 701955,
|
||||
"status": "sleeping"
|
||||
},
|
||||
{
|
||||
"cmdline": [
|
||||
"/home/frank14f/optane/MeltMind/artifacts/gre24-reproduction-2026-11-15/environment/venv/bin/python",
|
||||
"tools/phase2_h2_s2_guard.py",
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-2026-11-17",
|
||||
"4"
|
||||
],
|
||||
"ppid": 701941,
|
||||
"pid": 701956,
|
||||
"status": "sleeping"
|
||||
},
|
||||
{
|
||||
"cmdline": [
|
||||
"/home/frank14f/optane/MeltMind/artifacts/gre24-reproduction-2026-11-15/environment/venv/bin/python",
|
||||
"tools/phase2_h2_s2_worker.py",
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-2026-11-17",
|
||||
"1"
|
||||
],
|
||||
"ppid": 701953,
|
||||
"pid": 702009,
|
||||
"status": "sleeping"
|
||||
},
|
||||
{
|
||||
"cmdline": [
|
||||
"/home/frank14f/optane/MeltMind/artifacts/gre24-reproduction-2026-11-15/environment/venv/bin/python",
|
||||
"tools/phase2_h2_s2_worker.py",
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-2026-11-17",
|
||||
"3"
|
||||
],
|
||||
"ppid": 701955,
|
||||
"pid": 702016,
|
||||
"status": "sleeping"
|
||||
},
|
||||
{
|
||||
"cmdline": [
|
||||
"/home/frank14f/optane/MeltMind/artifacts/gre24-reproduction-2026-11-15/environment/venv/bin/python",
|
||||
"tools/phase2_h2_s2_worker.py",
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-2026-11-17",
|
||||
"4"
|
||||
],
|
||||
"ppid": 701956,
|
||||
"pid": 702018,
|
||||
"status": "running"
|
||||
},
|
||||
{
|
||||
"cmdline": [
|
||||
"/home/frank14f/optane/MeltMind/artifacts/gre24-reproduction-2026-11-15/environment/venv/bin/python",
|
||||
"tools/phase2_h2_s2_worker.py",
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-2026-11-17",
|
||||
"2"
|
||||
],
|
||||
"ppid": 701954,
|
||||
"pid": 702019,
|
||||
"status": "running"
|
||||
}
|
||||
],
|
||||
"code_pins_verified": true,
|
||||
"code_pin_count": 29,
|
||||
"bad": [],
|
||||
"contract_sha256": "76e1976bed88fc390812238c3d80693446daea3840dc009ac7d8870ddc43a31d",
|
||||
"service": "MainPID=701941\nActiveState=active\nSubState=running\n"
|
||||
},
|
||||
"phase2-gate-outcome-calibration-v2-2026-11-17": {
|
||||
"root": "/home/frank14f/MeltMind-runs/phase2-gate-outcome-calibration-v2-2026-11-17",
|
||||
"live": {
|
||||
"stage": "GATE_DEPENDENT_FEATURES_ONLY",
|
||||
"pending": 160,
|
||||
"active": [
|
||||
{
|
||||
"slot": 0,
|
||||
"pid": 721692,
|
||||
"index": 332
|
||||
}
|
||||
],
|
||||
"terminal": 332,
|
||||
"cumulative_cpu_seconds": 66935.52348795901,
|
||||
"unix": 1794927194.8617837
|
||||
},
|
||||
"complete_present": true,
|
||||
"processes": [],
|
||||
"code_pins_verified": true,
|
||||
"code_pin_count": 31,
|
||||
"bad": [],
|
||||
"contract_sha256": "5ea5b5bd06f9c40e57bf4f992a805ba42980223961ac1adbc94814e8f6473c35",
|
||||
"service": "MainPID=0\nActiveState=failed\nSubState=failed\n"
|
||||
}
|
||||
}
|
||||
},
|
||||
"latest_gate_terminal_at_S2_postfit_delivery": {
|
||||
"unix": 1794927358.4009368,
|
||||
"source": "/home/frank14f/MeltMind-runs/phase2-gate-outcome-calibration-v2-2026-11-17/complete.json",
|
||||
"source_sha256": "1846a533a86a585129ae867af4345000bc22f88f93e663605f7765263bed0f60",
|
||||
"terminal": "STOPPED",
|
||||
"reason": "RuntimeError: Full per-part reservation exceeds remaining budget",
|
||||
"terminal_records": 333,
|
||||
"statuses": {
|
||||
"PASS": 333
|
||||
},
|
||||
"never_attempted": 160,
|
||||
"cumulative_cpu_seconds": 66936.020178167,
|
||||
"accounting_complete": true,
|
||||
"service_MainPID": 0,
|
||||
"service_ExecMainStatus": 1,
|
||||
"original_PID_absent": true,
|
||||
"slack_correction": "1794927329.359199",
|
||||
"continuation_launched": false
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,19 @@
|
||||
{
|
||||
"unix": 1794927358.4009368,
|
||||
"source": "/home/frank14f/MeltMind-runs/phase2-gate-outcome-calibration-v2-2026-11-17/complete.json",
|
||||
"source_sha256": "1846a533a86a585129ae867af4345000bc22f88f93e663605f7765263bed0f60",
|
||||
"terminal": "STOPPED",
|
||||
"reason": "RuntimeError: Full per-part reservation exceeds remaining budget",
|
||||
"terminal_records": 333,
|
||||
"statuses": {
|
||||
"PASS": 333
|
||||
},
|
||||
"never_attempted": 160,
|
||||
"cumulative_cpu_seconds": 66936.020178167,
|
||||
"accounting_complete": true,
|
||||
"service_MainPID": 0,
|
||||
"service_ExecMainStatus": 1,
|
||||
"original_PID_absent": true,
|
||||
"slack_correction": "1794927329.359199",
|
||||
"continuation_launched": false
|
||||
}
|
||||
@@ -0,0 +1,188 @@
|
||||
{
|
||||
"unix": 1794927271.6056442,
|
||||
"status": "PREPARED_BINDING_DELIVERED_REVIEW_PENDING",
|
||||
"root": "/home/frank14f/MeltMind-runs/phase2-h2-s2-postfit-2026-11-17",
|
||||
"binding_sha256": "cabb17224f3a66c37f2e3089515802105194ad5d1a964e70b4c8ac355ac81e5a",
|
||||
"binding": {
|
||||
"reviewed_commit": "25e79995c7e414a0bb6bc66967e95d60289dc337",
|
||||
"contract_sha256": "0bd03f42c3b2db5bef26708b2ef873edfe674f724b59bc4b4de8641980d8d330",
|
||||
"implementation_proof_sha256": "c53fb8fb7bd134b2da4b6921b3770dc8fa0ade6cd7c67fb606ef981f8c2b93aa",
|
||||
"comparison_contract_sha256": "ea9616faebc17f66c4ce678fbc1358ed2292c423a0b016cc05693ebe52d2ecdf"
|
||||
},
|
||||
"code_files": 63,
|
||||
"synthetic_tests": 23,
|
||||
"review_request": "1794927225.586049",
|
||||
"review_request_link": "https://arcticnimbus.slack.com/archives/C0C78UFKGRM/p1794927225586049",
|
||||
"postfit_started": false,
|
||||
"new_review_present": false,
|
||||
"old_queues": {
|
||||
"phase2-h2-s2-2026-11-17": {
|
||||
"root": "/home/frank14f/MeltMind-runs/phase2-h2-s2-2026-11-17",
|
||||
"live": {
|
||||
"stage": "S2_TRAINING",
|
||||
"unix": 1794927271.289092,
|
||||
"active": [
|
||||
{
|
||||
"fold": 1,
|
||||
"guard_pid": 701953
|
||||
},
|
||||
{
|
||||
"fold": 2,
|
||||
"guard_pid": 701954
|
||||
},
|
||||
{
|
||||
"fold": 3,
|
||||
"guard_pid": 701955
|
||||
},
|
||||
{
|
||||
"fold": 4,
|
||||
"guard_pid": 701956
|
||||
}
|
||||
],
|
||||
"terminal": [],
|
||||
"new_gpu_seconds_charged": 8821.600681224954,
|
||||
"cumulative_h2_gpu_seconds": 27679.815589939943,
|
||||
"prediction_gpu_reserve": 2400,
|
||||
"train_gpu_cap": 122741.78509128501,
|
||||
"epoch_counts": {
|
||||
"1": 20,
|
||||
"2": 20,
|
||||
"3": 25,
|
||||
"4": 20
|
||||
},
|
||||
"rss_bytes": 7660138496,
|
||||
"disk_bytes": 20602880
|
||||
},
|
||||
"complete_present": false,
|
||||
"processes": [
|
||||
{
|
||||
"cmdline": [
|
||||
"/home/frank14f/optane/MeltMind/artifacts/gre24-reproduction-2026-11-15/environment/venv/bin/python",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s2_queue.py",
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-2026-11-17"
|
||||
],
|
||||
"ppid": 2234,
|
||||
"pid": 701941,
|
||||
"status": "sleeping"
|
||||
},
|
||||
{
|
||||
"cmdline": [
|
||||
"/home/frank14f/optane/MeltMind/artifacts/gre24-reproduction-2026-11-15/environment/venv/bin/python",
|
||||
"tools/phase2_h2_s2_guard.py",
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-2026-11-17",
|
||||
"1"
|
||||
],
|
||||
"ppid": 701941,
|
||||
"pid": 701953,
|
||||
"status": "sleeping"
|
||||
},
|
||||
{
|
||||
"cmdline": [
|
||||
"/home/frank14f/optane/MeltMind/artifacts/gre24-reproduction-2026-11-15/environment/venv/bin/python",
|
||||
"tools/phase2_h2_s2_guard.py",
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-2026-11-17",
|
||||
"2"
|
||||
],
|
||||
"ppid": 701941,
|
||||
"pid": 701954,
|
||||
"status": "sleeping"
|
||||
},
|
||||
{
|
||||
"cmdline": [
|
||||
"/home/frank14f/optane/MeltMind/artifacts/gre24-reproduction-2026-11-15/environment/venv/bin/python",
|
||||
"tools/phase2_h2_s2_guard.py",
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-2026-11-17",
|
||||
"3"
|
||||
],
|
||||
"ppid": 701941,
|
||||
"pid": 701955,
|
||||
"status": "sleeping"
|
||||
},
|
||||
{
|
||||
"cmdline": [
|
||||
"/home/frank14f/optane/MeltMind/artifacts/gre24-reproduction-2026-11-15/environment/venv/bin/python",
|
||||
"tools/phase2_h2_s2_guard.py",
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-2026-11-17",
|
||||
"4"
|
||||
],
|
||||
"ppid": 701941,
|
||||
"pid": 701956,
|
||||
"status": "sleeping"
|
||||
},
|
||||
{
|
||||
"cmdline": [
|
||||
"/home/frank14f/optane/MeltMind/artifacts/gre24-reproduction-2026-11-15/environment/venv/bin/python",
|
||||
"tools/phase2_h2_s2_worker.py",
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-2026-11-17",
|
||||
"1"
|
||||
],
|
||||
"ppid": 701953,
|
||||
"pid": 702009,
|
||||
"status": "sleeping"
|
||||
},
|
||||
{
|
||||
"cmdline": [
|
||||
"/home/frank14f/optane/MeltMind/artifacts/gre24-reproduction-2026-11-15/environment/venv/bin/python",
|
||||
"tools/phase2_h2_s2_worker.py",
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-2026-11-17",
|
||||
"3"
|
||||
],
|
||||
"ppid": 701955,
|
||||
"pid": 702016,
|
||||
"status": "sleeping"
|
||||
},
|
||||
{
|
||||
"cmdline": [
|
||||
"/home/frank14f/optane/MeltMind/artifacts/gre24-reproduction-2026-11-15/environment/venv/bin/python",
|
||||
"tools/phase2_h2_s2_worker.py",
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-2026-11-17",
|
||||
"4"
|
||||
],
|
||||
"ppid": 701956,
|
||||
"pid": 702018,
|
||||
"status": "running"
|
||||
},
|
||||
{
|
||||
"cmdline": [
|
||||
"/home/frank14f/optane/MeltMind/artifacts/gre24-reproduction-2026-11-15/environment/venv/bin/python",
|
||||
"tools/phase2_h2_s2_worker.py",
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-2026-11-17",
|
||||
"2"
|
||||
],
|
||||
"ppid": 701954,
|
||||
"pid": 702019,
|
||||
"status": "running"
|
||||
}
|
||||
],
|
||||
"code_pins_verified": true,
|
||||
"code_pin_count": 29,
|
||||
"bad": [],
|
||||
"contract_sha256": "76e1976bed88fc390812238c3d80693446daea3840dc009ac7d8870ddc43a31d",
|
||||
"service": "MainPID=701941\nActiveState=active\nSubState=running\n"
|
||||
},
|
||||
"phase2-gate-outcome-calibration-v2-2026-11-17": {
|
||||
"root": "/home/frank14f/MeltMind-runs/phase2-gate-outcome-calibration-v2-2026-11-17",
|
||||
"live": {
|
||||
"stage": "GATE_DEPENDENT_FEATURES_ONLY",
|
||||
"pending": 160,
|
||||
"active": [
|
||||
{
|
||||
"slot": 0,
|
||||
"pid": 721692,
|
||||
"index": 332
|
||||
}
|
||||
],
|
||||
"terminal": 332,
|
||||
"cumulative_cpu_seconds": 66935.52348795901,
|
||||
"unix": 1794927194.8617837
|
||||
},
|
||||
"complete_present": true,
|
||||
"processes": [],
|
||||
"code_pins_verified": true,
|
||||
"code_pin_count": 31,
|
||||
"bad": [],
|
||||
"contract_sha256": "5ea5b5bd06f9c40e57bf4f992a805ba42980223961ac1adbc94814e8f6473c35",
|
||||
"service": "MainPID=0\nActiveState=failed\nSubState=failed\n"
|
||||
}
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1 @@
|
||||
{"root": "/home/frank14f/MeltMind-runs/phase2-h2-s2-postfit-2026-11-17", "implementation": "25e79995c7e414a0bb6bc66967e95d60289dc337", "contract_sha256": "0bd03f42c3b2db5bef26708b2ef873edfe674f724b59bc4b4de8641980d8d330", "binding_sha256": "cabb17224f3a66c37f2e3089515802105194ad5d1a964e70b4c8ac355ac81e5a", "code_files": 63, "review_pending": true, "launched": false}
|
||||
@@ -468,3 +468,15 @@ service meltmind-h2-s2-20261117 / coordinator701941,guards701953/701954/701955
|
||||
|
||||
implementation1f57bad6/contract76e1976b/remote508ea39d,5HTTP/hash已核。首门通过不是200epoch终态,继续原每10epoch预算刷新;全四foldcheckpointfreeze后停止待新pooled scorer exactPASS,不复用S1review/不复跑fold0/data-sufficiency不启动。gate原666520只读321terminal/2active/170pending、64549.80548118CPU-s,原72000/68400及pins未改,95000接续条件保留未执行。下方首门pending快照为历史。
|
||||
|
||||
|
||||
|
||||
## H2-S2 pooled postfit已固定并交新语义审阅(2026-11-17)
|
||||
|
||||
Claude1794925470要求已完成:唯一准备根`/home/frank14f/MeltMind-runs/phase2-h2-s2-postfit-2026-11-17`,implementation25e79995c7e414a0bb6bc66967e95d60289dc337,contract0bd03f42c3b2db5bef26708b2ef873edfe674f724b59bc4b4de8641980d8d330,binding SHA cabb17224f3a66c37f2e3089515802105194ad5d1a964e70b4c8ac355ac81e5a;63源码固定、23CPU合成PASS。已在原Slack1794927225.586049交TRON→MeltAssist新pre-score审阅;实际PASS尚未收到,无后处理/等待器/新训练启动。报告docs/phase2-h2-s2-postfit-2026-11-17.md,审计artifacts/phase2-h2-s2-postfit-preparation-2026-11-17/live-audit.json。
|
||||
|
||||
fold0原121预测与指标hash复用;四新fold必须全部PASS、models/terminal冻结、原进程退出再接收四checkpoint。精确新PASS不能复用S1;全部552预测冻结再读431新test目标。552/204组件等权、2000/2026111408、H2−A field upper95CI<0唯一确认;B13/last/tau报告,负值不裁剪。原2400预测GPU-s/3600postfitCPU-s/总144000旧账一次,0重训/重试/新阶段。
|
||||
|
||||
本次只读实核原训练stage=S2_TRAINING、epoch={'1': 20, '2': 20, '3': 25, '4': 20}、累计27679.815590GPU-s;gate 332terminal/1active/160pending、66935.523488CPU-s。两原service及源码pins核验通过,不改原72k/68400,不提前启动95k接续。先前准备/运行快照是历史;后续读取真实终态,不能因idle重启。
|
||||
|
||||
|
||||
只读终态补正:gate原666520于22:53:15已退出,service MainPID0/ExecMainStatus1。实际STOPPED,完整单件预算预留不足,非数值失败;333终态,状态{'PASS': 333},160never-attempted,CPU66936.020178167、accounting_complete=true。原始live332/1active是退出前历史。已在Slack1794927329.359199纠正;95k一次标记接续尚未启动,不能重启原队列。详见gate-terminal-audit.json。
|
||||
@@ -0,0 +1,29 @@
|
||||
# H2 S2 pooled postfit: exact review preparation
|
||||
|
||||
Claude1794925470.101149 authorizes preparing this entry concurrently with the original four-fold training. This document describes implementation, not executed predictions or a received independent PASS. The preparation ends by delivering the exact binding to MeltAssist. It launches no waiter, prediction, scoring, fitting or new stage.
|
||||
|
||||
The unique prepared root is `/home/frank14f/MeltMind-runs/phase2-h2-s2-postfit-2026-11-17`. Original training remains `/home/frank14f/MeltMind-runs/phase2-h2-s2-2026-11-17`, implementation1f57bad66eb062d4f2269347532321929f873717, contract76e1976bed88fc390812238c3d80693446daea3840dc009ac7d8870ddc43a31d. Neither its code nor the gate outcome queue is changed.
|
||||
|
||||
## Exact inputs and execution barriers
|
||||
|
||||
Reuse all121 original S1 fold0 prediction entries and per-part metrics by hash. No fold0 prediction, target-array read or per-part metric recomputation. All five folds retain the original H1/A learning/test roles, baseline model hashes, learning-only clocks, eligible query identities and multiplicities. B13-E1b predictions and per-part metrics are frozen references; its42 invalid-feature parts stay in the552-part roster.
|
||||
|
||||
Before any new prediction, require all four original training terminal entries PASS/exit0, exact settled ledger, original models and terminal Git freezes, absence of training processes, final epoch200 checkpoints and each fold's learning-only normalization. The future checkpoint hashes are adopted only from those verified original freezes, never guessed during preparation. The new adoption freeze binds all four checkpoints. Reject partial training, failed guards, altered checkpoints, wrong folds/steps, validation-selected checkpoints and doubled old costs.
|
||||
|
||||
The new `expected-review-binding.json` identifies the implementation commit, contract SHA256, implementation-proof SHA256, comparison-contract SHA256 and complete code SHA256 map. An actual MeltAssist PASS matching all five fields must be copied and frozen as `implementation-review`; the shared validator must pass. S1's PASS cannot satisfy this gate. Prediction and scoring each revalidate it; scoring does so before parsing old score values or reading test target arrays. All552 predictions are hash-frozen before scoring. Only431 folds1–4 targets are then read from the original caches; no new decoding.
|
||||
|
||||
## Predictions and statistics
|
||||
|
||||
Four sequential full-graph predictions use the original fold GPUs, FP32 H2 model, original FP64 F skip, frozen final checkpoints and normalization. Query extraction follows the full-node prediction. Negative values remain unchanged; full-node and eligible unique-query negative counts are reported separately. No clipping or baseline refitting.
|
||||
|
||||
Pool552 parts /204 components with unchanged E1b `pooled()`: component means weighted equally, per-arm components require all constituent part metrics defined, paired comparison uses the intersection. Use2000 paired component bootstrap replicates, seed2026111408. CONFIRMED requires complete paired field support204/204 and a finite upper95%CI for H2−A strictly below0. Incomplete field support is PARK. H2−B13-E1b, last-fill location/Rg and Kendall tau are report-only; per-fold results also reported. Undefined metrics are retained, not zero-filled or silently dropped within components. This is recipe confirmation after observing fold0, not an untouched-five-fold preregistration or causal identification of connectivity.
|
||||
|
||||
## Bounded accounting and failure preservation
|
||||
|
||||
Total H2 cap remains144000 GPU-seconds. Original S1 settled18858.21490871499 is counted once in the original fourfold training terminal ledger; that full settled cumulative amount is adopted once. New prediction allowance2400 GPU-seconds includes four600-second folds and all their idle/setup/failure time. Each fold has a conservative60-second setup charge, three idle polls, original GPU identity, and no retry. Predictions run sequentially;15GiB recorded/device VRAM,16GiB tree RSS and20GiB total postfit output caps apply. Failed workers retain peak records and cause a stop.
|
||||
|
||||
Postfit cap3600 CPU-seconds includes360 conservative setup (120 preparation plus four60 local allowances), worker/child CPU and coordinator cost. Before each fold reserve540 worker CPU-seconds for each remaining prediction,600 for scoring and120 for terminal work. Actual child CPU and explicitly recorded worker CPU are combined by maximum, not added twice. Score worker has600 CPU-seconds /780 wall-seconds /16GiB RSS. No resource or numerical failure triggers a retry. Failures, original S1 evidence and training pins remain intact. No data-sufficiency, extra training or other follow-on stage is launched.
|
||||
|
||||
## Validation status
|
||||
|
||||
23 CPU synthetic checks passed: exact actual Git/file review validation, all-four terminal and ledger identities, checkpoint fold/final epoch rules, remaining resource reservations, failure peak/coverage checks, strict confirmation boundaries, component weighting and undefined-component semantics, exact fold0 metric/manifest reuse,552-entry uniqueness and target-read ordering. No real target values, new predictions or GPU workload were used. Independent semantic review remains pending until its actual matching receipt is received.
|
||||
@@ -1,3 +1,13 @@
|
||||
## H2-S2 pooled postfit已固定并交新语义审阅(2026-11-17)
|
||||
|
||||
Claude1794925470要求已完成:唯一准备根`/home/frank14f/MeltMind-runs/phase2-h2-s2-postfit-2026-11-17`,implementation25e79995c7e414a0bb6bc66967e95d60289dc337,contract0bd03f42c3b2db5bef26708b2ef873edfe674f724b59bc4b4de8641980d8d330,binding SHA cabb17224f3a66c37f2e3089515802105194ad5d1a964e70b4c8ac355ac81e5a;63源码固定、23CPU合成PASS。已在原Slack1794927225.586049交TRON→MeltAssist新pre-score审阅;实际PASS尚未收到,无后处理/等待器/新训练启动。报告docs/phase2-h2-s2-postfit-2026-11-17.md,审计artifacts/phase2-h2-s2-postfit-preparation-2026-11-17/live-audit.json。
|
||||
|
||||
fold0原121预测与指标hash复用;四新fold必须全部PASS、models/terminal冻结、原进程退出再接收四checkpoint。精确新PASS不能复用S1;全部552预测冻结再读431新test目标。552/204组件等权、2000/2026111408、H2−A field upper95CI<0唯一确认;B13/last/tau报告,负值不裁剪。原2400预测GPU-s/3600postfitCPU-s/总144000旧账一次,0重训/重试/新阶段。
|
||||
|
||||
本次只读实核原训练stage=S2_TRAINING、epoch={'1': 20, '2': 20, '3': 25, '4': 20}、累计27679.815590GPU-s;gate 332terminal/1active/160pending、66935.523488CPU-s。训练service活跃、两原源码pins核验通过;gate服务已退出见下方补正,不改原72k/68400,不提前启动95k接续。先前准备/运行快照是历史;后续读取真实终态,不能因idle重启。
|
||||
|
||||
只读终态补正:gate原666520于22:53:15已退出,service MainPID0/ExecMainStatus1。实际STOPPED,完整单件预算预留不足,非数值失败;333终态,状态{'PASS': 333},160never-attempted,CPU66936.020178167、accounting_complete=true。原始live332/1active是退出前历史。已在Slack1794927329.359199纠正;95k一次标记接续尚未启动,不能重启原队列。详见gate-terminal-audit.json。
|
||||
|
||||
## H2-S2首完整四fold成本门PASS,原队列已继续(2026-11-17 22:22)
|
||||
|
||||
首epoch实际102.782503/107.558575/89.451310/107.500166s,原1.25safety、旧账一次和2400prediction reserve后总投影124682.378895535<144000GPU-s;四单job投影均<30685.446帽。原唯一coordinator701941自动解锁,实核四foldepoch2/steps664/651/566/621,9expectedprocesses、pinsPASS、无globalstop。首门及不可变001epoch/normalization/身份/启动审计固定e767d5777c4d3bc0a16018113ac95a00f24938c7;不得重开、重试或改pins。实际运行根`/home/frank14f/MeltMind-runs/phase2-h2-s2-2026-11-17`。新报告docs/phase2-h2-s2-actual-start-2026-11-17.md。
|
||||
|
||||
@@ -0,0 +1,10 @@
|
||||
{
|
||||
"status": "PASS",
|
||||
"tests": 23,
|
||||
"elapsed_seconds": 1.477,
|
||||
"log_sha256": "c1dd4f98c9903679cc62a5b942d241dfa2638c082964efb88e8cc5b304f26bda",
|
||||
"test_source_sha256": "7f15ff9ceded979ea7a991ddd9d1abb143eac29fa1bf02d848b0ece8aba1c31e",
|
||||
"scope": "CPU synthetic plus actual isolated Git/file review-validator tests; no real prediction, target read, fitting or CUDA workload",
|
||||
"setup_cpu_seconds_reserved": 360,
|
||||
"unix": 1794927191.9851246
|
||||
}
|
||||
@@ -0,0 +1,100 @@
|
||||
{
|
||||
"approval": "1794925470.101149",
|
||||
"test_parts": 552,
|
||||
"test_components": 204,
|
||||
"fold0_prediction_and_metric_hash_reuse": true,
|
||||
"new_prediction_parts": 431,
|
||||
"checkpoint_folds": [
|
||||
1,
|
||||
2,
|
||||
3,
|
||||
4
|
||||
],
|
||||
"checkpoint_epoch": 200,
|
||||
"seed": 0,
|
||||
"A_models": [
|
||||
{
|
||||
"path": "/home/frank14f/optane/MeltMind/artifacts/r4bc-execution-2026-11-16/jobs/job-0065/output/model.joblib",
|
||||
"sha256": "5f8914daf0f775f9d7986ffe180d8005f8bcf799ff768866ad36997f79fded7f",
|
||||
"format": "joblib"
|
||||
},
|
||||
{
|
||||
"path": "/home/frank14f/optane/MeltMind/artifacts/r4bc-execution-2026-11-16/jobs/job-0132/output/model.joblib",
|
||||
"sha256": "f422f81e299fa5e03e2ddf34a6d493dfbdbea59ff2e4c9a1ca9ebfb125b3f852",
|
||||
"format": "joblib"
|
||||
},
|
||||
{
|
||||
"path": "/home/frank14f/optane/MeltMind/artifacts/r4bc-execution-2026-11-16/jobs/job-0199/output/model.joblib",
|
||||
"sha256": "155257c6c2bc000f01467df19ca070e8698f7ea1a1f17bb80b057aeee7ea178a",
|
||||
"format": "joblib"
|
||||
},
|
||||
{
|
||||
"path": "/home/frank14f/optane/MeltMind/artifacts/r4bc-execution-2026-11-16/jobs/job-0266/output/model.joblib",
|
||||
"sha256": "e0304b935308a7131279c743581c8625d2ecad693f50fb32e7dbb19bb4f50951",
|
||||
"format": "joblib"
|
||||
},
|
||||
{
|
||||
"path": "/home/frank14f/optane/MeltMind/artifacts/r4bc-execution-2026-11-16/jobs/job-0333/output/model.joblib",
|
||||
"sha256": "37d460545972b18bd21538e22806c300ff77c676e5123d1baf723cd343d0e071",
|
||||
"format": "joblib"
|
||||
}
|
||||
],
|
||||
"clocks": [
|
||||
{
|
||||
"fold": 0,
|
||||
"c": 1.026602777749673,
|
||||
"path": "/home/frank14f/optane/MeltMind/artifacts/phase2-h1-execution-2026-11-17/fits/job-00/clock.json",
|
||||
"sha256": "c7b76bd79e17565b05669ec31a0e9fbf16e0113e6d6a3d913514075e5aedea96"
|
||||
},
|
||||
{
|
||||
"fold": 1,
|
||||
"c": 1.0302015087964238,
|
||||
"path": "/home/frank14f/optane/MeltMind/artifacts/phase2-h1-execution-2026-11-17/fits/job-01/clock.json",
|
||||
"sha256": "2d56271a8f03bf131f405581c2a9bf5a1c75512384efca963c970ae907a014ef"
|
||||
},
|
||||
{
|
||||
"fold": 2,
|
||||
"c": 1.0257851343108513,
|
||||
"path": "/home/frank14f/optane/MeltMind/artifacts/phase2-h1-execution-2026-11-17/fits/job-02/clock.json",
|
||||
"sha256": "852b856f9b4b4459a54ed66bd5e26637eacbcb8161b2c2a360d3593a478d9911"
|
||||
},
|
||||
{
|
||||
"fold": 3,
|
||||
"c": 1.0272648651548586,
|
||||
"path": "/home/frank14f/optane/MeltMind/artifacts/phase2-h1-execution-2026-11-17/fits/job-03/clock.json",
|
||||
"sha256": "464aa534cdf92f31709ddaae1aa8715fd24d6540c3d1f3935197a027adc50221"
|
||||
},
|
||||
{
|
||||
"fold": 4,
|
||||
"c": 1.0300213434206609,
|
||||
"path": "/home/frank14f/optane/MeltMind/artifacts/phase2-h1-execution-2026-11-17/fits/job-04/clock.json",
|
||||
"sha256": "657c97876089317c71fb0d0bf750490040e4d19492c189737d5801fa53d9e88f"
|
||||
}
|
||||
],
|
||||
"metrics": [
|
||||
"relative_l2",
|
||||
"last_location",
|
||||
"kendall_tau_b"
|
||||
],
|
||||
"query": "Original eligible S2 queries; multiplicities once; raw-ID minimum tie rule; last distance/Rg",
|
||||
"pooling": "Original E1b pooled(): per-arm component requires all constituent part metrics defined; component means equally weighted; paired intersection",
|
||||
"bootstrap": {
|
||||
"replicates": 2000,
|
||||
"seed": 2026111408
|
||||
},
|
||||
"confirmation": "CONFIRMED iff complete552/204 paired field H2-A upper95CI<0; strict finite bound; incomplete field support PARK",
|
||||
"report_only": [
|
||||
"H2-B13_E1b",
|
||||
"last_location",
|
||||
"kendall_tau_b"
|
||||
],
|
||||
"negative_clipping": false,
|
||||
"all_four_checkpoints_frozen_and_training_exited_before_prediction": true,
|
||||
"actual_new_semantic_PASS_before_any_target": true,
|
||||
"predictions_frozen_before_score": true,
|
||||
"new_fits": 0,
|
||||
"new_decodes": 0,
|
||||
"no_retry": true,
|
||||
"no_follow_on_launch": true,
|
||||
"interpretation": "Five-fold recipe confirmation after observing fold0; not an untouched-five-fold preregistration or connectivity causal claim"
|
||||
}
|
||||
File diff suppressed because it is too large.
Load diff
@@ -0,0 +1,71 @@
|
||||
{
|
||||
"reviewed_commit": "25e79995c7e414a0bb6bc66967e95d60289dc337",
|
||||
"contract_sha256": "0bd03f42c3b2db5bef26708b2ef873edfe674f724b59bc4b4de8641980d8d330",
|
||||
"implementation_proof_sha256": "c53fb8fb7bd134b2da4b6921b3770dc8fa0ade6cd7c67fb606ef981f8c2b93aa",
|
||||
"comparison_contract_sha256": "ea9616faebc17f66c4ce678fbc1358ed2292c423a0b016cc05693ebe52d2ecdf",
|
||||
"code_sha256": {
|
||||
"/home/frank14f/optane/MeltMind/tools/test_phase2_h2_s2.py": "4235890fd194b7778e937f0be9245b964512df2c75d962ef703e3a988c5b8805",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_main_memory_guard.py": "e094f5443e11518efe9018215c11718a8e18aa462a977bf43e974b2d0b2aebb1",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_recovery_gpu_supervisor.py": "689c0f3179e4fcd362e546d90214829a9e4f8739abdbdf74b560763e5a392f15",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_recovery_monitoring.py": "442c2a9c75e8c00b30a8f2e91cb52d87c466b19a80aea9128b10c021e61b922e",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_precision_continuation_guard.py": "01235d897c7c03e5ffa109c58b14f5d24d128e8f1ba873e1f76cfb14e02f7a7f",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_gpu_idle_gate.py": "e2b1613e7e6ea398a8d555220a9365a1ad998ab4ef4b2bf8ec7cc4e31f8b4b9f",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_pilot_train.py": "e9cceead51a5cc092aad23409ad98909c84a62e7e10571e750d1e9be1ad3b948",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_models/network.py": "5137e845d81d1223afc13bf87e03ccba1132a3baa75d5189a8f320b2752be071",
|
||||
"/home/frank14f/optane/MeltMind/tools/freeze_g1_components.py": "d0cee60d95243ac2c59ef3f2244ba818315718c4e3d1895f2f9f064f1f7f5ecd",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h1_worker_v2.py": "1543df85b615bb69b79bacd572074a7c77ae19dab110197601573a7e5805ed20",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_residual_math.py": "ffbd9624dec79bdaabf95244fe87209055908f31a977e73ea4377045d6f8fe62",
|
||||
"/home/frank14f/optane/MeltMind/tools/gre24_models.py": "53b6fee8b700b9b982b06d29908ce2ce23e0c57512c46dbf74b2aade79e3ac3a",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s2_guard.py": "6bc7624b0cfe850591f77153cd2034eefdd62f7e63b93211ccdb87eb8aba346c",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s2_common.py": "308486507ae3589a0e6cc40565f6e60e6d2f63ae74c4d828bdacf33b3b8b8062",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_home_freeze_io.py": "bf09204b85ed2a7038acdd202257f618474d1110a9d6d772f105a3ffb334d3ff",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_local_freeze_io.py": "60adf3382577613c5234f5cf30779336b9951b66e90c075cb7cd0adab9cc42dd",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_fixed_io.py": "ea17fdfc7508f3f980e3cdf69534b1eb8ceb8a8e469c32451b2d5e16d0ada575",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s2_worker.py": "37dfd75fff50b3d65f4ef37f4255536616582fe5f3d01155a0e037325ab5c7cd",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_pilot_worker.py": "14651ca402e6d02466d8afdb77bc9e62fb2e752f0c0c3d20d98593157d5ebe81",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_model.py": "fab2470dd0654ca74c3b7d893f35ad1f3ae6c5c08ffe334ef6312cf6ec87892a",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_data.py": "7e03dd7a94e7bb43593277f183c222473d7cb7aaf2f60e53b45ab8d9499fcb09",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s1_common.py": "38a29a34663a5d6310571eeff53b5d18a0355e51b9fa58a08033f64fa379a9df",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s1_worker.py": "ecee04b183ef2b93b7b39a447d48c40a5843d3190e48d2d7eb4939811a5db9b0",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s2_prepare.py": "964fae13ac95e4de4dc50b3c0a80791d86349ce3a9f3c5a6e7dc3cf8ca041fad",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_prepare.py": "3a31af0766f628a80f38e0b7286beb11ca3109188eabc4adfdb2ecb0eef3dc27",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s2_queue.py": "d82dfb365956e08250758bb4c651a19fbecce92fe209072f5f8ed35970450735",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s1_launch.py": "480fac25dea5a2006bb84130a4b0a0b253b3040b2150a56c38eada62b2c259e6",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s1_guard.py": "ae94711adc17a7630f398e8d4120fa3cc8f60979a00122c4cd2dc364c382ced5",
|
||||
"/home/frank14f/optane/MeltMind/docs/phase2-h2-s2-execution-contract-2026-11-17.md": "1e8e5ef31650519737cd6cd247abc0bd04744dd93ffb013eff88866d42f03073",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_b13_e1b.py": "e0fb6c9ecb51fdae24706e792edccb5d2259082f06aa2e292e7a23772328330d",
|
||||
"/home/frank14f/optane/MeltMind/tools/r4bc_common.py": "27543fff987975ece767167f9771e6621c825035a053805604e4a696794dfcbd",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_recovery_supervisor.py": "f563ca3c2665f5baf0c04f296616e4640423a59b815fbb00479af0f29e57a919",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_models/features.py": "3b34ae61cf8012ebd58b7505a73d7aa43d8ee1890ac54f98d7917b67887a45a4",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_custody_statistics.py": "188565801f6901d0192f79aec0889e632501e672a4149d76d1835efbb9b45447",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_models/objectives.py": "5b6527a45f08690055d572335470ce5979e8bb86e1cab00de594f79ae21316b1",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_custody_common.py": "588c646101be0d50f505f9a12a1c98361aa3437c8c594eecb7321831f717a741",
|
||||
"/home/frank14f/optane/MeltMind/tools/r4bc_guard.py": "437a756ddf10b125dce450f463e76e28df52724f8c5766e0a989dfa536805628",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_b1213_gate.py": "d5a023749e9d8bb6e04a2a2cbdbb8f0d9568924c534f36102d06707859853e10",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_b1213_features.py": "d8e975b0a5362be7ff00fbf554f2b438b54af1b6b95f8993e7c2beae77f2c8f2",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_b12_math.py": "5465b34be5578cc62a7afc67e3cee3cd489ebb75ce6622a3fa320d1b81196c57",
|
||||
"/home/frank14f/optane/MeltMind/tools/o1_v2.py": "7bfc81d820a1be7d40b97b2c0b65a616812cb7615b99d80c4ac589788282c291",
|
||||
"/home/frank14f/optane/MeltMind/tools/o1_partition.py": "5e7dbf97604b9ea9646113a9847e7232020e5f9248fad388177c32ae2a7117da",
|
||||
"/home/frank14f/optane/MeltMind/tools/o1_prepare.py": "1a695477ee6605d75e7e991205e3e6a49a0907a7733491af8d2771c7bc6a5c23",
|
||||
"/home/frank14f/optane/MeltMind/tools/gre24_features.py": "92bca4d348f31246c2e931949e2ca38103881b1b784a621a42044db56a460026",
|
||||
"/home/frank14f/optane/MeltMind/tools/boundary_distance.py": "a45c1cbce92521a96eb814913b4f465b926c0b3883932ab0b7c37d11db113ed7",
|
||||
"/home/frank14f/optane/MeltMind/tools/audit_samples.py": "a3daf70ec3bcb77bf46d9dc4887d552cf7c422117880a169969780012278cd70",
|
||||
"/home/frank14f/optane/MeltMind/tools/gre24_features_v2.py": "7dcd5ac218ffdc1ca149edcd87e8658fb3c59e18c0fb9ce47985c7f1357d2cdf",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s1_test_data.py": "fc0770fe7de50692dd0ce97ad3868a72fd1b65ab339ebdb9b9460adb51632473",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s1_postfit.py": "767621616b2584cf43be9eaf03bb71e92a7bb5e9e120d323bd8e1c1a7e7966ba",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s1_prediction_guard.py": "5642c166c6627dbbf6fc529454c33c868d887acc60f08fdfa3b9fb26d44f675b",
|
||||
"/home/frank14f/optane/MeltMind/tools/prepare_phase2_h2_s1_postfit.py": "977d4df66f007a089be328c190f58b3525d9bcaa6cf54bd6279d0966ca7ae874",
|
||||
"/home/frank14f/optane/MeltMind/tools/test_phase2_h2_s1_postfit.py": "29c0cfbcfe32cc9b9bc394de7991bd1896c6b544983609ebf247d93170ce1e25",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_gate_outcome_review.py": "915617aeb8292d03c654e637bc34a014c254fd178a773f135d1806c4b41372a8",
|
||||
"/home/frank14f/optane/MeltMind/tools/test_phase2_gate_outcome_postfit.py": "ae3f0849b8cf0abba75d3a7fd7b14d756234a4c094c23e446c5aaa61b75d7c79",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_gate_outcome_statistics.py": "8e9f781cc4975beff245d991a9b886e0d8e7bf6d66c0ff0337d76dccaddda8e1",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_gate_outcome_postfit.py": "b05ee93cfac1f032c84b5a0e33e6d052a4f3b507ad6c717aa6d42456ae4bc512",
|
||||
"/home/frank14f/optane/MeltMind/docs/phase2-h2-s1-postfit-2026-11-17.md": "abe0bf6a1baee7554432fc209c6b049a948a6db3d4ba705b33fbd5a0059410b0",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s2_postfit_common.py": "e4616411da44af7bb0a5e8f4e3645754e722f06bb4386054569017c3c4496f81",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s2_postfit.py": "a6d413d61327b19880e13ebe1bd9511d6d8a21c7858bcd7c66ce03cb72e86861",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s2_prediction_guard.py": "98b12df1690fc580ad0f986e6d6f077a6860af840455d720d6608314fd640785",
|
||||
"/home/frank14f/optane/MeltMind/tools/prepare_phase2_h2_s2_postfit.py": "c48c981295b16c4a4892237964b05de2b9f5d3ecf485fa94d1c473e0e894f9ff",
|
||||
"/home/frank14f/optane/MeltMind/tools/test_phase2_h2_s2_postfit.py": "7f15ff9ceded979ea7a991ddd9d1abb143eac29fa1bf02d848b0ece8aba1c31e",
|
||||
"/home/frank14f/optane/MeltMind/docs/phase2-h2-s2-postfit-2026-11-17.md": "4b0dcf0b9cb63f910d90acc42659d9ce8e727d956b1397c881104edadfe12158"
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,4 @@
|
||||
{
|
||||
"contract_sha256": "0bd03f42c3b2db5bef26708b2ef873edfe674f724b59bc4b4de8641980d8d330",
|
||||
"proof_sha256": "c53fb8fb7bd134b2da4b6921b3770dc8fa0ade6cd7c67fb606ef981f8c2b93aa"
|
||||
}
|
||||
@@ -0,0 +1,145 @@
|
||||
{
|
||||
"commit": "25e79995c7e414a0bb6bc66967e95d60289dc337",
|
||||
"files": {
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-postfit-2026-11-17/contract.json": "0bd03f42c3b2db5bef26708b2ef873edfe674f724b59bc4b4de8641980d8d330",
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-postfit-2026-11-17/comparison-contract.json": "ea9616faebc17f66c4ce678fbc1358ed2292c423a0b016cc05693ebe52d2ecdf",
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-postfit-2026-11-17/acceptance.json": "88ce3612e605cfcb228773cffd5ad1343eb19c5b41d5459fa5237e208ad9b92a",
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-postfit-2026-11-17/synthetic-attempt1.log": "c1dd4f98c9903679cc62a5b942d241dfa2638c082964efb88e8cc5b304f26bda",
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-postfit-2026-11-17/preparation-audit.json": "2f79b06ef8503a2dce3e7d3ecae6a8972c12a8a0fc577a4279f78713de12b5d2",
|
||||
"/home/frank14f/optane/MeltMind/tools/test_phase2_h2_s2.py": "4235890fd194b7778e937f0be9245b964512df2c75d962ef703e3a988c5b8805",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_main_memory_guard.py": "e094f5443e11518efe9018215c11718a8e18aa462a977bf43e974b2d0b2aebb1",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_recovery_gpu_supervisor.py": "689c0f3179e4fcd362e546d90214829a9e4f8739abdbdf74b560763e5a392f15",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_recovery_monitoring.py": "442c2a9c75e8c00b30a8f2e91cb52d87c466b19a80aea9128b10c021e61b922e",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_precision_continuation_guard.py": "01235d897c7c03e5ffa109c58b14f5d24d128e8f1ba873e1f76cfb14e02f7a7f",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_gpu_idle_gate.py": "e2b1613e7e6ea398a8d555220a9365a1ad998ab4ef4b2bf8ec7cc4e31f8b4b9f",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_pilot_train.py": "e9cceead51a5cc092aad23409ad98909c84a62e7e10571e750d1e9be1ad3b948",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_models/network.py": "5137e845d81d1223afc13bf87e03ccba1132a3baa75d5189a8f320b2752be071",
|
||||
"/home/frank14f/optane/MeltMind/tools/freeze_g1_components.py": "d0cee60d95243ac2c59ef3f2244ba818315718c4e3d1895f2f9f064f1f7f5ecd",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h1_worker_v2.py": "1543df85b615bb69b79bacd572074a7c77ae19dab110197601573a7e5805ed20",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_residual_math.py": "ffbd9624dec79bdaabf95244fe87209055908f31a977e73ea4377045d6f8fe62",
|
||||
"/home/frank14f/optane/MeltMind/tools/gre24_models.py": "53b6fee8b700b9b982b06d29908ce2ce23e0c57512c46dbf74b2aade79e3ac3a",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s2_guard.py": "6bc7624b0cfe850591f77153cd2034eefdd62f7e63b93211ccdb87eb8aba346c",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s2_common.py": "308486507ae3589a0e6cc40565f6e60e6d2f63ae74c4d828bdacf33b3b8b8062",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_home_freeze_io.py": "bf09204b85ed2a7038acdd202257f618474d1110a9d6d772f105a3ffb334d3ff",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_local_freeze_io.py": "60adf3382577613c5234f5cf30779336b9951b66e90c075cb7cd0adab9cc42dd",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_fixed_io.py": "ea17fdfc7508f3f980e3cdf69534b1eb8ceb8a8e469c32451b2d5e16d0ada575",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s2_worker.py": "37dfd75fff50b3d65f4ef37f4255536616582fe5f3d01155a0e037325ab5c7cd",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_pilot_worker.py": "14651ca402e6d02466d8afdb77bc9e62fb2e752f0c0c3d20d98593157d5ebe81",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_model.py": "fab2470dd0654ca74c3b7d893f35ad1f3ae6c5c08ffe334ef6312cf6ec87892a",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_data.py": "7e03dd7a94e7bb43593277f183c222473d7cb7aaf2f60e53b45ab8d9499fcb09",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s1_common.py": "38a29a34663a5d6310571eeff53b5d18a0355e51b9fa58a08033f64fa379a9df",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s1_worker.py": "ecee04b183ef2b93b7b39a447d48c40a5843d3190e48d2d7eb4939811a5db9b0",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s2_prepare.py": "964fae13ac95e4de4dc50b3c0a80791d86349ce3a9f3c5a6e7dc3cf8ca041fad",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_prepare.py": "3a31af0766f628a80f38e0b7286beb11ca3109188eabc4adfdb2ecb0eef3dc27",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s2_queue.py": "d82dfb365956e08250758bb4c651a19fbecce92fe209072f5f8ed35970450735",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s1_launch.py": "480fac25dea5a2006bb84130a4b0a0b253b3040b2150a56c38eada62b2c259e6",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s1_guard.py": "ae94711adc17a7630f398e8d4120fa3cc8f60979a00122c4cd2dc364c382ced5",
|
||||
"/home/frank14f/optane/MeltMind/docs/phase2-h2-s2-execution-contract-2026-11-17.md": "1e8e5ef31650519737cd6cd247abc0bd04744dd93ffb013eff88866d42f03073",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_b13_e1b.py": "e0fb6c9ecb51fdae24706e792edccb5d2259082f06aa2e292e7a23772328330d",
|
||||
"/home/frank14f/optane/MeltMind/tools/r4bc_common.py": "27543fff987975ece767167f9771e6621c825035a053805604e4a696794dfcbd",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_recovery_supervisor.py": "f563ca3c2665f5baf0c04f296616e4640423a59b815fbb00479af0f29e57a919",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_models/features.py": "3b34ae61cf8012ebd58b7505a73d7aa43d8ee1890ac54f98d7917b67887a45a4",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_custody_statistics.py": "188565801f6901d0192f79aec0889e632501e672a4149d76d1835efbb9b45447",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_models/objectives.py": "5b6527a45f08690055d572335470ce5979e8bb86e1cab00de594f79ae21316b1",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_custody_common.py": "588c646101be0d50f505f9a12a1c98361aa3437c8c594eecb7321831f717a741",
|
||||
"/home/frank14f/optane/MeltMind/tools/r4bc_guard.py": "437a756ddf10b125dce450f463e76e28df52724f8c5766e0a989dfa536805628",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_b1213_gate.py": "d5a023749e9d8bb6e04a2a2cbdbb8f0d9568924c534f36102d06707859853e10",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_b1213_features.py": "d8e975b0a5362be7ff00fbf554f2b438b54af1b6b95f8993e7c2beae77f2c8f2",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_b12_math.py": "5465b34be5578cc62a7afc67e3cee3cd489ebb75ce6622a3fa320d1b81196c57",
|
||||
"/home/frank14f/optane/MeltMind/tools/o1_v2.py": "7bfc81d820a1be7d40b97b2c0b65a616812cb7615b99d80c4ac589788282c291",
|
||||
"/home/frank14f/optane/MeltMind/tools/o1_partition.py": "5e7dbf97604b9ea9646113a9847e7232020e5f9248fad388177c32ae2a7117da",
|
||||
"/home/frank14f/optane/MeltMind/tools/o1_prepare.py": "1a695477ee6605d75e7e991205e3e6a49a0907a7733491af8d2771c7bc6a5c23",
|
||||
"/home/frank14f/optane/MeltMind/tools/gre24_features.py": "92bca4d348f31246c2e931949e2ca38103881b1b784a621a42044db56a460026",
|
||||
"/home/frank14f/optane/MeltMind/tools/boundary_distance.py": "a45c1cbce92521a96eb814913b4f465b926c0b3883932ab0b7c37d11db113ed7",
|
||||
"/home/frank14f/optane/MeltMind/tools/audit_samples.py": "a3daf70ec3bcb77bf46d9dc4887d552cf7c422117880a169969780012278cd70",
|
||||
"/home/frank14f/optane/MeltMind/tools/gre24_features_v2.py": "7dcd5ac218ffdc1ca149edcd87e8658fb3c59e18c0fb9ce47985c7f1357d2cdf",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s1_test_data.py": "fc0770fe7de50692dd0ce97ad3868a72fd1b65ab339ebdb9b9460adb51632473",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s1_postfit.py": "767621616b2584cf43be9eaf03bb71e92a7bb5e9e120d323bd8e1c1a7e7966ba",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s1_prediction_guard.py": "5642c166c6627dbbf6fc529454c33c868d887acc60f08fdfa3b9fb26d44f675b",
|
||||
"/home/frank14f/optane/MeltMind/tools/prepare_phase2_h2_s1_postfit.py": "977d4df66f007a089be328c190f58b3525d9bcaa6cf54bd6279d0966ca7ae874",
|
||||
"/home/frank14f/optane/MeltMind/tools/test_phase2_h2_s1_postfit.py": "29c0cfbcfe32cc9b9bc394de7991bd1896c6b544983609ebf247d93170ce1e25",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_gate_outcome_review.py": "915617aeb8292d03c654e637bc34a014c254fd178a773f135d1806c4b41372a8",
|
||||
"/home/frank14f/optane/MeltMind/tools/test_phase2_gate_outcome_postfit.py": "ae3f0849b8cf0abba75d3a7fd7b14d756234a4c094c23e446c5aaa61b75d7c79",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_gate_outcome_statistics.py": "8e9f781cc4975beff245d991a9b886e0d8e7bf6d66c0ff0337d76dccaddda8e1",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_gate_outcome_postfit.py": "b05ee93cfac1f032c84b5a0e33e6d052a4f3b507ad6c717aa6d42456ae4bc512",
|
||||
"/home/frank14f/optane/MeltMind/docs/phase2-h2-s1-postfit-2026-11-17.md": "abe0bf6a1baee7554432fc209c6b049a948a6db3d4ba705b33fbd5a0059410b0",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s2_postfit_common.py": "e4616411da44af7bb0a5e8f4e3645754e722f06bb4386054569017c3c4496f81",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s2_postfit.py": "a6d413d61327b19880e13ebe1bd9511d6d8a21c7858bcd7c66ce03cb72e86861",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s2_prediction_guard.py": "98b12df1690fc580ad0f986e6d6f077a6860af840455d720d6608314fd640785",
|
||||
"/home/frank14f/optane/MeltMind/tools/prepare_phase2_h2_s2_postfit.py": "c48c981295b16c4a4892237964b05de2b9f5d3ecf485fa94d1c473e0e894f9ff",
|
||||
"/home/frank14f/optane/MeltMind/tools/test_phase2_h2_s2_postfit.py": "7f15ff9ceded979ea7a991ddd9d1abb143eac29fa1bf02d848b0ece8aba1c31e",
|
||||
"/home/frank14f/optane/MeltMind/docs/phase2-h2-s2-postfit-2026-11-17.md": "4b0dcf0b9cb63f910d90acc42659d9ce8e727d956b1397c881104edadfe12158"
|
||||
},
|
||||
"git_paths": {
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-postfit-2026-11-17/contract.json": "home-runs/phase2-h2-s2-postfit-2026-11-17/contract.json",
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-postfit-2026-11-17/comparison-contract.json": "home-runs/phase2-h2-s2-postfit-2026-11-17/comparison-contract.json",
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-postfit-2026-11-17/acceptance.json": "home-runs/phase2-h2-s2-postfit-2026-11-17/acceptance.json",
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-postfit-2026-11-17/synthetic-attempt1.log": "home-runs/phase2-h2-s2-postfit-2026-11-17/synthetic-attempt1.log",
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-postfit-2026-11-17/preparation-audit.json": "home-runs/phase2-h2-s2-postfit-2026-11-17/preparation-audit.json",
|
||||
"/home/frank14f/optane/MeltMind/tools/test_phase2_h2_s2.py": "tools/test_phase2_h2_s2.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_main_memory_guard.py": "tools/s2_main_memory_guard.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_recovery_gpu_supervisor.py": "tools/s2_recovery_gpu_supervisor.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_recovery_monitoring.py": "tools/s2_recovery_monitoring.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_precision_continuation_guard.py": "tools/s2_precision_continuation_guard.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_gpu_idle_gate.py": "tools/s2_gpu_idle_gate.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_pilot_train.py": "tools/s2_pilot_train.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_models/network.py": "tools/s2_models/network.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/freeze_g1_components.py": "tools/freeze_g1_components.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h1_worker_v2.py": "tools/phase2_h1_worker_v2.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_residual_math.py": "tools/phase2_residual_math.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/gre24_models.py": "tools/gre24_models.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s2_guard.py": "tools/phase2_h2_s2_guard.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s2_common.py": "tools/phase2_h2_s2_common.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_home_freeze_io.py": "tools/phase2_home_freeze_io.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_local_freeze_io.py": "tools/phase2_local_freeze_io.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_fixed_io.py": "tools/phase2_fixed_io.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s2_worker.py": "tools/phase2_h2_s2_worker.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_pilot_worker.py": "tools/phase2_h2_pilot_worker.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_model.py": "tools/phase2_h2_model.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_data.py": "tools/phase2_h2_data.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s1_common.py": "tools/phase2_h2_s1_common.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s1_worker.py": "tools/phase2_h2_s1_worker.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s2_prepare.py": "tools/phase2_h2_s2_prepare.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_prepare.py": "tools/phase2_h2_prepare.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s2_queue.py": "tools/phase2_h2_s2_queue.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s1_launch.py": "tools/phase2_h2_s1_launch.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s1_guard.py": "tools/phase2_h2_s1_guard.py",
|
||||
"/home/frank14f/optane/MeltMind/docs/phase2-h2-s2-execution-contract-2026-11-17.md": "docs/phase2-h2-s2-execution-contract-2026-11-17.md",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_b13_e1b.py": "tools/phase2_b13_e1b.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/r4bc_common.py": "tools/r4bc_common.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_recovery_supervisor.py": "tools/s2_recovery_supervisor.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_models/features.py": "tools/s2_models/features.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_custody_statistics.py": "tools/s2_custody_statistics.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_models/objectives.py": "tools/s2_models/objectives.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/s2_custody_common.py": "tools/s2_custody_common.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/r4bc_guard.py": "tools/r4bc_guard.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_b1213_gate.py": "tools/phase2_b1213_gate.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_b1213_features.py": "tools/phase2_b1213_features.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_b12_math.py": "tools/phase2_b12_math.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/o1_v2.py": "tools/o1_v2.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/o1_partition.py": "tools/o1_partition.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/o1_prepare.py": "tools/o1_prepare.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/gre24_features.py": "tools/gre24_features.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/boundary_distance.py": "tools/boundary_distance.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/audit_samples.py": "tools/audit_samples.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/gre24_features_v2.py": "tools/gre24_features_v2.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s1_test_data.py": "tools/phase2_h2_s1_test_data.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s1_postfit.py": "tools/phase2_h2_s1_postfit.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s1_prediction_guard.py": "tools/phase2_h2_s1_prediction_guard.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/prepare_phase2_h2_s1_postfit.py": "tools/prepare_phase2_h2_s1_postfit.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/test_phase2_h2_s1_postfit.py": "tools/test_phase2_h2_s1_postfit.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_gate_outcome_review.py": "tools/phase2_gate_outcome_review.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/test_phase2_gate_outcome_postfit.py": "tools/test_phase2_gate_outcome_postfit.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_gate_outcome_statistics.py": "tools/phase2_gate_outcome_statistics.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_gate_outcome_postfit.py": "tools/phase2_gate_outcome_postfit.py",
|
||||
"/home/frank14f/optane/MeltMind/docs/phase2-h2-s1-postfit-2026-11-17.md": "docs/phase2-h2-s1-postfit-2026-11-17.md",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s2_postfit_common.py": "tools/phase2_h2_s2_postfit_common.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s2_postfit.py": "tools/phase2_h2_s2_postfit.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/phase2_h2_s2_prediction_guard.py": "tools/phase2_h2_s2_prediction_guard.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/prepare_phase2_h2_s2_postfit.py": "tools/prepare_phase2_h2_s2_postfit.py",
|
||||
"/home/frank14f/optane/MeltMind/tools/test_phase2_h2_s2_postfit.py": "tools/test_phase2_h2_s2_postfit.py",
|
||||
"/home/frank14f/optane/MeltMind/docs/phase2-h2-s2-postfit-2026-11-17.md": "docs/phase2-h2-s2-postfit-2026-11-17.md"
|
||||
},
|
||||
"remote_published": false,
|
||||
"local_repository": "/tmp/meltmind-phase2-local-fixed-20261117"
|
||||
}
|
||||
@@ -0,0 +1,27 @@
|
||||
{
|
||||
"status": "PREPARED_NOT_LAUNCHED",
|
||||
"unix": 1794927191.9852877,
|
||||
"training_implementation": "1f57bad66eb062d4f2269347532321929f873717",
|
||||
"S1_prediction_commit": "b4f488cc6ec25b9935545aa7663ee5dc8ebfd05c",
|
||||
"S1_terminal_commit": "71e8d5a1f809e1c00548843f147386c13df0fea0",
|
||||
"test_parts": 552,
|
||||
"test_components": 204,
|
||||
"fold_test_counts": [
|
||||
121,
|
||||
91,
|
||||
99,
|
||||
119,
|
||||
122
|
||||
],
|
||||
"original_learning_test_roles_and_A_models_equal": true,
|
||||
"original_prediction_hashes_checked": 673,
|
||||
"new_fits": 0,
|
||||
"new_predictions": 0,
|
||||
"new_scores": 0,
|
||||
"old_score_values_read": false,
|
||||
"test_target_values_read": false,
|
||||
"training_untouched": true,
|
||||
"gate_queue_untouched": true,
|
||||
"future_checkpoint_adoption": "Not present or required at preparation; all four original models/terminal freezes and process exit mandatory at execution",
|
||||
"actual_new_review_pending": true
|
||||
}
|
||||
@@ -0,0 +1,4 @@
|
||||
{
|
||||
"contract_sha256": "0bd03f42c3b2db5bef26708b2ef873edfe674f724b59bc4b4de8641980d8d330",
|
||||
"proof_sha256": "00063dae3816eafd0530230b06ae5f0a21832a93770a0289b7da9ea2b50d1ed9"
|
||||
}
|
||||
@@ -0,0 +1,17 @@
|
||||
{
|
||||
"commit": "b2462a1c6b8f85dcd3483f3f5fc47a92ca05d799",
|
||||
"files": {
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-postfit-2026-11-17/contract.json": "0bd03f42c3b2db5bef26708b2ef873edfe674f724b59bc4b4de8641980d8d330",
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-postfit-2026-11-17/expected-review-binding.json": "cabb17224f3a66c37f2e3089515802105194ad5d1a964e70b4c8ac355ac81e5a",
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-postfit-2026-11-17/implementation-git-freeze.json": "c53fb8fb7bd134b2da4b6921b3770dc8fa0ade6cd7c67fb606ef981f8c2b93aa",
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-postfit-2026-11-17/implementation-authorization.json": "ce55382804f5197c92e5b6babba58de68fb499cb4fcb8459f99f5a6834b27210"
|
||||
},
|
||||
"git_paths": {
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-postfit-2026-11-17/contract.json": "home-runs/phase2-h2-s2-postfit-2026-11-17/contract.json",
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-postfit-2026-11-17/expected-review-binding.json": "home-runs/phase2-h2-s2-postfit-2026-11-17/expected-review-binding.json",
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-postfit-2026-11-17/implementation-git-freeze.json": "home-runs/phase2-h2-s2-postfit-2026-11-17/implementation-git-freeze.json",
|
||||
"/home/frank14f/MeltMind-runs/phase2-h2-s2-postfit-2026-11-17/implementation-authorization.json": "home-runs/phase2-h2-s2-postfit-2026-11-17/implementation-authorization.json"
|
||||
},
|
||||
"remote_published": false,
|
||||
"local_repository": "/tmp/meltmind-phase2-local-fixed-20261117"
|
||||
}
|
||||
@@ -0,0 +1,28 @@
|
||||
test_complete_roster_and_exact_fold0_entries (tools.test_phase2_h2_s2_postfit.ManifestReuse) ... ok
|
||||
test_duplicate_missing_and_wrong_new_fold_path_rejected (tools.test_phase2_h2_s2_postfit.ManifestReuse) ... ok
|
||||
test_fold0_changed_hash_rejected (tools.test_phase2_h2_s2_postfit.ManifestReuse) ... ok
|
||||
test_negative_counts_not_silently_changed_or_boolean (tools.test_phase2_h2_s2_postfit.ManifestReuse) ... ok
|
||||
test_actual_frozen_pass (test_phase2_gate_outcome_postfit.SemanticReview) ... ok
|
||||
test_changed_comparison_rejected (test_phase2_gate_outcome_postfit.SemanticReview) ... ok
|
||||
test_changed_source_rejected (test_phase2_gate_outcome_postfit.SemanticReview) ... ok
|
||||
test_fail_missing_and_unrelated_receipts (test_phase2_gate_outcome_postfit.SemanticReview) ... ok
|
||||
test_review_omitted_from_freeze (test_phase2_gate_outcome_postfit.SemanticReview) ... ok
|
||||
test_wrong_exact_identities (test_phase2_gate_outcome_postfit.SemanticReview) ... ok
|
||||
test_component_equal_weights_and_b13_report_comparison (tools.test_phase2_h2_s2_postfit.StatisticsAndCustody) ... ok
|
||||
test_fold0_metrics_are_reused_not_recomputed (tools.test_phase2_h2_s2_postfit.StatisticsAndCustody) ... ok
|
||||
test_fold0_prediction_forbidden_before_config_or_CUDA (tools.test_phase2_h2_s2_postfit.StatisticsAndCustody) ... ok
|
||||
test_only_strict_complete_field_vs_A_confirms (tools.test_phase2_h2_s2_postfit.StatisticsAndCustody) ... ok
|
||||
test_original_whole_component_undefined_rule (tools.test_phase2_h2_s2_postfit.StatisticsAndCustody) ... ok
|
||||
test_semantic_pass_and_all_models_precede_value_reads (tools.test_phase2_h2_s2_postfit.StatisticsAndCustody) ... ok
|
||||
test_all_four_costs_and_prior_once (tools.test_phase2_h2_s2_postfit.TerminalAndBudget) ... ok
|
||||
test_budget_reserves_all_remaining_predictions_and_score (tools.test_phase2_h2_s2_postfit.TerminalAndBudget) ... ok
|
||||
test_correct_new_fold_final_checkpoint_only (tools.test_phase2_h2_s2_postfit.TerminalAndBudget) ... ok
|
||||
test_mismatched_or_double_counted_ledger_rejected (tools.test_phase2_h2_s2_postfit.TerminalAndBudget) ... ok
|
||||
test_partial_failed_or_wrong_role_models_rejected (tools.test_phase2_h2_s2_postfit.TerminalAndBudget) ... ok
|
||||
test_prediction_final_peaks_and_per_fold_coverage (tools.test_phase2_h2_s2_postfit.TerminalAndBudget) ... ok
|
||||
test_still_training_without_model_freeze_cannot_enter (tools.test_phase2_h2_s2_postfit.TerminalAndBudget) ... ok
|
||||
|
||||
----------------------------------------------------------------------
|
||||
Ran 23 tests in 1.477s
|
||||
|
||||
OK
|
||||
@@ -0,0 +1,125 @@
|
||||
"""Single S2 postfit: four frozen-checkpoint predictions then exact-reviewed pooled score."""
|
||||
import argparse,fcntl,os,resource,sys,time
|
||||
from pathlib import Path
|
||||
import numpy as np
|
||||
import psutil
|
||||
from phase2_h1_worker_v2 import save,sha,checked
|
||||
from phase2_home_freeze_io import freeze,verify
|
||||
from phase2_gate_outcome_review import validate_review
|
||||
from phase2_h2_s2_postfit_common import read,config,terminal_inputs,adopted_inputs,checkpoint_identity,comparisons,decision,require_budget,roster,reuse_fold0_metric
|
||||
from phase2_h2_s2_common import fold_view
|
||||
from phase2_h2_s1_postfit import query_index
|
||||
from phase2_residual_math import decode_seconds
|
||||
from s2_custody_common import prediction_inputs
|
||||
from phase2_b1213_features import cached
|
||||
from s2_custody_statistics import part_metrics
|
||||
from r4bc_guard import supervise
|
||||
|
||||
def predict(root,fold):
|
||||
import torch
|
||||
from phase2_h2_model import H2Geo
|
||||
from phase2_h2_data import to_device,arguments
|
||||
from phase2_h2_s1_test_data import load_test_part
|
||||
if fold not in (1,2,3,4):raise PermissionError('Fold0 predictions are hash reuse only')
|
||||
c,pilot,tc=config(root);validate_review(root);adopted=adopted_inputs(root,c,pilot,tc);m=next(m for m in adopted['models'] if m['fold']==fold);view=fold_view(pilot,fold);f=view['folds'][0];job=tc['jobs'][fold-1];stats=read(m['standardization']);by=roster(c,pilot)
|
||||
dest=root/'folds'/str(fold)/'predictions';dest.mkdir(exist_ok=False)
|
||||
if torch.__version__!=pilot['versions']['torch'] or np.__version__!=pilot['versions']['numpy']:raise ValueError('Original scientific library versions')
|
||||
if os.environ.get('CUDA_VISIBLE_DEVICES')!=job['gpu_uuid'] or os.environ.get('PYTORCH_CUDA_ALLOC_CONF')!='expandable_segments:True':raise ValueError('GPU/allocator identity')
|
||||
os.sched_setaffinity(0,job['worker_cpus']);torch.set_num_threads(1);torch.set_num_interop_threads(1);torch.backends.cuda.matmul.allow_tf32=False;torch.backends.cudnn.allow_tf32=False
|
||||
torch.cuda.set_device(0);torch.cuda.init();torch.cuda.reset_peak_memory_stats();start=time.monotonic();cpu=time.process_time();entries=[]
|
||||
def peaks():return dict(peak_allocated_bytes=int(torch.cuda.max_memory_allocated()),peak_reserved_bytes=int(torch.cuda.max_memory_reserved()))
|
||||
try:
|
||||
blob=torch.load(checked(m['checkpoint'],m['checkpoint_sha256']),map_location='cpu',weights_only=False);checkpoint_identity(blob,c,fold,len(f['learning']));model=H2Geo().cuda();model.load_state_dict(blob['model'],strict=True);model.eval();del blob
|
||||
save(dest/'environment.json',dict(fold=fold,torch=torch.__version__,gpu_uuid=job['gpu_uuid'],allocator_settings=str(torch.cuda.memory._snapshot().get('allocator_settings',{})),model_dtype='FP32',autocast=False,checkpoint_sha256=m['checkpoint_sha256']))
|
||||
with torch.no_grad():
|
||||
for case in f['test']:
|
||||
row=by[case];data=to_device(load_test_part(view,case,stats),'cuda');residual=model(*arguments(data),checkpoint_blocks=False)
|
||||
if not bool(torch.isfinite(residual).all()):raise ValueError('Nonfinite H2 residual')
|
||||
seconds=decode_seconds(residual.cpu().numpy(),data['F'],f['c'],data['t_set']);torch.cuda.synchronize();q=prediction_inputs(row);a=c['baseline_predictions'][case]
|
||||
with np.load(checked(a['path'],a['sha256']),allow_pickle=False) as old:
|
||||
ix=query_index(data['raw_ids'],q,old);A=old['A'].copy();G=old['pred'].copy()
|
||||
if not np.isfinite(A).all() or not np.isfinite(G).all() or len(seconds)!=row['nodes']:raise ValueError('Full-node/baseline coverage')
|
||||
path=dest/(case+'.npz');np.savez_compressed(path,seconds=seconds,raw_ids=data['raw_ids'],pred=seconds[ix],A=A,B13_E1b=G,query_raw_ids=q['raw_ids'],counts=q['counts'],coordinates_over_Rg=q['coordinates_over_Rg'],near_wall=q['near_wall'])
|
||||
entries.append(dict(case=case,component_id=row['component_id'],fold=fold,path=str(path),sha256=sha(path),nodes=row['nodes'],query_nodes=len(ix),A_source_sha256=a['sha256'],negative_full_predictions=int(np.sum(seconds<0)),negative_query_predictions=int(np.sum(seconds[ix]<0)),origin='S2_FROZEN_CHECKPOINT'))
|
||||
save(dest/'progress.json',dict(stage='TEST_PREDICTION_NO_TARGETS',fold=fold,parts=len(entries),wall_seconds=time.monotonic()-start,cpu_seconds=time.process_time()-cpu,**peaks()));del data,residual,seconds
|
||||
save(dest/'manifest.json',dict(entries=entries,test_targets_read=False,new_fits=0,full_nodes=sum(e['nodes'] for e in entries)))
|
||||
save(dest/'result.json',dict(status='COMPLETE',fold=fold,parts=len(entries),test_targets_read=False,wall_seconds=time.monotonic()-start,cpu_seconds=time.process_time()-cpu,**peaks()))
|
||||
except BaseException as exc:
|
||||
save(dest/'failure.json',dict(status='FAILED',fold=fold,reason=type(exc).__name__+': '+str(exc),completed_parts=len(entries),wall_seconds=time.monotonic()-start,cpu_seconds=time.process_time()-cpu,**peaks()));raise
|
||||
|
||||
def manifest_identity(entries,c,pilot):
|
||||
by=roster(c,pilot);old={e['case']:e for e in read(c['fold0_manifest'])['entries']}
|
||||
if len(entries)!=552 or len({e['case'] for e in entries})!=552 or {e['case'] for e in entries}!=set(by):raise ValueError('All552 distinct predictions required')
|
||||
for e in entries:
|
||||
row=by[e['case']];fold=next(a['fold'] for a in row['assignments'] if a['role']=='test')
|
||||
if e['fold']!=fold or e['component_id']!=row['component_id'] or e['nodes']!=row['nodes']:raise ValueError('Prediction roster identity')
|
||||
if fold==0:
|
||||
if e.get('origin')!='S1_HASH_REUSE' or {k:v for k,v in e.items() if k!='origin'}!=old[e['case']]:raise ValueError('Fold0 entry must be exact original hash reuse')
|
||||
elif e.get('origin')!='S2_FROZEN_CHECKPOINT' or e['A_source_sha256']!=c['baseline_predictions'][e['case']]['sha256'] or Path(e['path']).resolve()!=(Path(c['postfit_root'])/'folds'/str(fold)/'predictions'/(e['case']+'.npz')).resolve():raise ValueError('New prediction source identity')
|
||||
for key,limit in [('negative_full_predictions',e['nodes']),('negative_query_predictions',e['query_nodes'])]:
|
||||
if isinstance(e[key],bool) or not isinstance(e[key],int) or not 0<=e[key]<=limit:raise ValueError('Invalid unclipped negative count')
|
||||
return by
|
||||
|
||||
def score(root):
|
||||
c,pilot,tc=config(root);validate_review(root);adopted_inputs(root,c,pilot,tc);verify(root,'predictions')
|
||||
entries=read(root/'predictions-manifest.json')['entries'];by=manifest_identity(entries,c,pilot);dest=root/'score';dest.mkdir(exist_ok=False)
|
||||
save(dest/'review-consumed.json',dict(unix=time.time(),review_sha256=sha(root/'review.json'),before_old_score_or_target_values=True))
|
||||
oldA={r['case']:r for r in read(checked(c['baseline_score'],c['source_sha256'][c['baseline_score']]))['rows']};oldH={r['case']:r for r in read(checked(c['fold0_score'],c['source_sha256'][c['fold0_score']]))['rows']};rows=[];access=[]
|
||||
if len(oldH)!=121 or set(oldH)!=set(pilot['folds'][0]['test']) or len(oldA)!=552 or set(oldA)!=set(by):raise ValueError('Frozen reference score coverage')
|
||||
for e in entries:
|
||||
case=e['case'];row=by[case];old=oldA[case];checked(e['path'],e['sha256'])
|
||||
if old['fold']!=e['fold'] or old['component_id']!=e['component_id']:raise ValueError('Baseline score pairing')
|
||||
if e['fold']==0:
|
||||
metric=reuse_fold0_metric(oldH[case],old,e) # No fold0 target read, prediction, or per-part metric recomputation.
|
||||
else:
|
||||
access.append(dict(case=case,fold=e['fold'],role='test',target_sha256=row['v5_cache']['output_sha256']['seconds.npy']));save(dest/'target-access.json',access)
|
||||
truth=cached(row,'seconds');raw=cached(row,'raw_ids');base=c['baseline_predictions'][case]
|
||||
with np.load(checked(e['path'],e['sha256']),allow_pickle=False) as z,np.load(checked(base['path'],base['sha256']),allow_pickle=False) as original:
|
||||
q={k:z['query_raw_ids' if k=='raw_ids' else k] for k in ['raw_ids','counts','coordinates_over_Rg','near_wall']};ix=query_index(raw,q,original)
|
||||
if not np.array_equal(z['raw_ids'],raw) or not np.array_equal(z['pred'],z['seconds'][ix]) or not np.array_equal(z['A'],original['A']) or not np.array_equal(z['B13_E1b'],original['pred']):raise ValueError('Full/query/reference prediction identity')
|
||||
if len(z['seconds'])!=e['nodes'] or len(ix)!=e['query_nodes'] or int(np.sum(z['seconds']<0))!=e['negative_full_predictions'] or int(np.sum(z['pred']<0))!=e['negative_query_predictions']:raise ValueError('Negative counts or coverage differ')
|
||||
metric=part_metrics(z['pred'],truth[ix],q['counts'],q['near_wall'],q['raw_ids'],q['coordinates_over_Rg'],int(case.split('_')[0]))
|
||||
rows.append(dict(case=case,component_id=e['component_id'],fold=e['fold'],H2=metric,A=old['A'],B13_E1b=old['new'],baseline_G_valid=old['G_valid'],H2_origin=e['origin']));save(dest/'progress.json',dict(parts=len(rows),new_test_target_reads=len(access)))
|
||||
if len(access)!=431 or len({x['case'] for x in access})!=431:raise ValueError('Exactly431 new held-out targets; fold0 untouched')
|
||||
metrics=comparisons(rows);result=decision(metrics,len(rows),len({r['component_id'] for r in rows}))
|
||||
save(dest/'result.json',dict(status='COMPLETE',decision=result,parts=552,components=204,comparisons=metrics,per_fold={str(f):comparisons([r for r in rows if r['fold']==f]) for f in range(5)},rows=rows,confirmation_rule='pooled H2-A field CI upper<0, complete552/204; B13/last/tau report only',bootstrap=dict(replicates=2000,seed=2026111408),negative_predictions=dict(full_nodes=sum(e['negative_full_predictions'] for e in entries),eligible_unique_query_nodes=sum(e['negative_query_predictions'] for e in entries),clipped=False),full_nodes=sum(e['nodes'] for e in entries),eligible_unique_query_nodes=sum(e['query_nodes'] for e in entries),baseline_G_invalid_parts=sum(not r['baseline_G_valid'] for r in rows),fold0_score_reused=121,new_test_target_reads=431,new_decodes=0,new_fits=0,review_sha256=sha(root/'review.json'),data_sufficiency_started=False,interpretation='Five-fold recipe confirmation after observing fold0; no pure connectivity causality or untouched-five-fold preregistration claim'))
|
||||
|
||||
def run(root):
|
||||
start=time.process_time();u0=resource.getrusage(resource.RUSAGE_CHILDREN);c,pilot,tc=config(root);validate_review(root);adopted=terminal_inputs(c,pilot,tc);lock=(root/'run.lock').open('a');fcntl.flock(lock,fcntl.LOCK_EX|fcntl.LOCK_NB)
|
||||
if (root/'started.json').exists():raise FileExistsError('Single pooled postfit attempt')
|
||||
for p in psutil.process_iter(['pid','cmdline']):
|
||||
if p.pid!=os.getpid() and any(Path(a).name=='phase2_h2_s2_postfit.py' for a in p.info['cmdline'] or []):raise RuntimeError('Existing S2 postfit process')
|
||||
os.sched_setaffinity(0,[12,13]);os.environ['PYTORCH_CUDA_ALLOC_CONF']='expandable_segments:True';gpu=0.;charged=0.;records=[];prediction_start=None;prior=adopted['prior_gpu_seconds']
|
||||
def cpu():
|
||||
u=resource.getrusage(resource.RUSAGE_CHILDREN);return 360+time.process_time()-start+max(charged,u.ru_utime+u.ru_stime-u0.ru_utime-u0.ru_stime)
|
||||
save(root/'started.json',dict(pid=os.getpid(),unix=time.time(),prior_gpu_once=prior,old_s1_gpu_once=adopted['old_s1_gpu_once'],new_training_gpu_once=adopted['new_training_gpu_once'],review_sha256=sha(root/'review.json'),new_fits=0,fold0_reexecution=False))
|
||||
try:
|
||||
require_budget(prior,gpu,cpu(),4);save(root/'models-adopted.json',adopted);train=Path(c['training_root'])
|
||||
freeze(root,'models',[root/'contract.json',root/'models-adopted.json',train/'complete.json',train/'models-terminal.json',train/'models-git-freeze.json',train/'terminal-git-freeze.json',*[Path(m['checkpoint']) for m in adopted['models']],*[Path(m['standardization']) for m in adopted['models']]])
|
||||
entries=[dict(e,origin='S1_HASH_REUSE') for e in read(c['fold0_manifest'])['entries']]
|
||||
for e in entries:checked(e['path'],e['sha256'])
|
||||
from phase2_h2_s2_prediction_guard import run_prediction
|
||||
for fold in range(1,5):
|
||||
require_budget(prior,gpu,cpu(),5-fold);(root/'folds'/str(fold)).mkdir(parents=True,exist_ok=False);prediction_start=time.monotonic();receipt=run_prediction(root,c,pilot,fold,prior+gpu,600)
|
||||
observed=60+time.monotonic()-prediction_start;prediction_start=None;cost=max(receipt['gpu_seconds_charged'],observed);gpu+=cost;charged+=receipt['worker']['cpu_seconds'] if receipt['worker'] else 0.;records.append(dict(mode='predict',fold=fold,charged_gpu_seconds=cost,guard=receipt))
|
||||
save(root/'resource-progress.json',dict(records=records,prediction_gpu_seconds=gpu,postfit_cpu_seconds=cpu(),cumulative_h2_gpu_seconds=prior+gpu))
|
||||
if receipt['status']!='PASS' or cost>600 or gpu>2400 or prior+gpu>144000:raise RuntimeError('Prediction not PASS or cumulative GPU limit')
|
||||
m=read(root/'folds'/str(fold)/'predictions/manifest.json')
|
||||
if len(m['entries'])!=len(pilot['folds'][fold]['test']) or m['full_nodes']!=pilot['folds'][fold]['test_nodes']:raise ValueError('Full-fold prediction coverage')
|
||||
for e in m['entries']:checked(e['path'],e['sha256'])
|
||||
entries.extend(m['entries'])
|
||||
manifest_identity(entries,c,pilot);save(root/'predictions-manifest.json',dict(entries=entries,new_fits=0,fold0_reused=121,new_predictions=431,test_targets_read=False));freeze(root,'predictions',[root/'contract.json',root/'predictions-manifest.json',root/'models-adopted.json',root/'resource-progress.json',*[Path(p) for p in c['code_sha256']]])
|
||||
require_budget(prior,gpu,cpu(),0)
|
||||
g=supervise([sys.executable,'tools/phase2_h2_s2_postfit.py','score',str(root)],root/'score-guard',dict(cpu_seconds=600,wall_seconds=780,rss_bytes=16*2**30,disk_bytes=2*2**30));charged+=g['cpu_seconds'];records.append(dict(mode='score',guard=g))
|
||||
if g['status']!='PASS' or cpu()>3600:raise RuntimeError('Score not PASS or CPU limit')
|
||||
save(root/'complete.json',dict(status='COMPLETE',records=records,prediction_gpu_seconds=gpu,postfit_cpu_seconds=cpu(),prior_gpu_once=prior,cumulative_h2_gpu_seconds=prior+gpu,decision=read(root/'score/result.json')['decision'],new_fits=0,fold0_reexecution=False,data_sufficiency_started=False,no_retry=True));freeze(root,'terminal',[root/'contract.json',root/'complete.json',root/'score/result.json',root/'review.json'])
|
||||
save(root/'post-terminal-cost.json',dict(postfit_cpu_seconds=cpu(),prediction_gpu_seconds=gpu,cumulative_h2_gpu_seconds=prior+gpu,within_cap=cpu()<=3600 and gpu<=2400 and prior+gpu<=144000))
|
||||
except BaseException as exc:
|
||||
if prediction_start is not None:gpu+=60+time.monotonic()-prediction_start
|
||||
save(root/'complete.json',dict(status='STOPPED',reason=type(exc).__name__+': '+str(exc),records=records,postfit_cpu_seconds=cpu(),prediction_gpu_seconds=gpu,prior_gpu_once=prior,cumulative_h2_gpu_seconds=prior+gpu,new_fits=0,no_retry=True));raise
|
||||
if __name__=='__main__':
|
||||
p=argparse.ArgumentParser();p.add_argument('action',choices=['run','predict','score']);p.add_argument('root',type=Path);p.add_argument('fold',type=int,nargs='?');a=p.parse_args()
|
||||
if a.action=='predict':predict(a.root,a.fold)
|
||||
elif a.fold is not None:p.error('Fold allowed only for predict')
|
||||
elif a.action=='score':score(a.root)
|
||||
else:run(a.root)
|
||||
@@ -0,0 +1,103 @@
|
||||
"""S2 postfit identities, all-four terminal barrier and unchanged paired statistics."""
|
||||
import json,math,os
|
||||
from pathlib import Path
|
||||
import numpy as np
|
||||
from phase2_h1_worker_v2 import checked,sha
|
||||
from phase2_home_freeze_io import verify
|
||||
from phase2_h2_s2_common import confirmed
|
||||
from phase2_b13_e1b import pooled
|
||||
|
||||
def read(p):return json.loads(Path(p).read_text())
|
||||
def roster(c,pilot):
|
||||
rows=c['rows'];by={r['case']:r for r in rows}
|
||||
if len(rows)!=552 or len(by)!=552 or len({r['component_id'] for r in rows})!=204:raise ValueError('Fixed552/204 roster')
|
||||
alltest=[];component_folds={}
|
||||
for f in pilot['folds']:
|
||||
for case in f['test']:
|
||||
row=by[case];pr=pilot['parts'][case];a=c['baseline_predictions'][case]
|
||||
tests=[v['fold'] for v in row['assignments'] if v['role']=='test']
|
||||
if tests!=[f['fold']] or a['fold']!=f['fold'] or row['component_id']!=a['component_id'] or row['component_id']!=pr['clock_group'] or row['v5_cache']['output_sha256']!=pr['output_sha256'] or row['nodes']!=pr['nodes'] or row['preparation']['t_set']!=pr['t_set']:raise ValueError('Original test/cache/clock/query baseline identity')
|
||||
group=row['component_id']
|
||||
if group in component_folds and component_folds[group]!=f['fold']:raise ValueError('Component spans test folds')
|
||||
component_folds[group]=f['fold'];alltest.append(case)
|
||||
if len(alltest)!=552 or len(set(alltest))!=552 or set(alltest)!=set(by):raise ValueError('Original five-fold partition')
|
||||
return by
|
||||
|
||||
def config(root):
|
||||
c=read(root/'contract.json');verify(root,'implementation')
|
||||
if c['approval']!='1794925470.101149' or c['total_h2_gpu_seconds']!=144000 or c['prediction_gpu_seconds']!=2400 or c['postfit_cpu_seconds']!=3600 or c['new_fits']!=0 or c['setup_cpu_seconds']!=360:raise PermissionError('Fixed pooled postfit scope')
|
||||
for p,h in {**c['source_sha256'],**c['code_sha256']}.items():checked(p,h)
|
||||
train=Path(c['training_root']);proof=verify(train,'implementation')
|
||||
if proof['commit']!=c['training_implementation']:raise ValueError('Wrong frozen training implementation')
|
||||
tc=read(train/'contract.json');pilot=read(c['pilot_contract'])
|
||||
if np.__version__!=pilot['versions']['numpy']:raise ValueError('Original NumPy version required')
|
||||
verify(Path(c['s1_root']),'predictions');verify(Path(c['s1_root']),'terminal')
|
||||
if tc['prior_gpu_seconds']!=18858.21490871499 or tc['prediction_gpu_reserve']!=2400 or tc['postfit_cpu_cap']!=3600 or tc['epochs']!=200 or tc['seed']!=0:raise ValueError('Original training budget/recipe')
|
||||
if [j['fold'] for j in tc['jobs']]!=[1,2,3,4] or c['training_jobs']!=tc['jobs']:raise ValueError('Original four jobs')
|
||||
roster(c,pilot);return c,pilot,tc
|
||||
|
||||
def terminal_identity(tc,done,ledger,models):
|
||||
if done.get('status')!='TRAIN_COMPLETE_POSTFIT_PENDING' or done.get('postfit_started') is not False:raise PermissionError('All-four training terminal required')
|
||||
entries=done.get('terminal',[])
|
||||
if len(entries)!=4 or {e['fold'] for e in entries}!={1,2,3,4} or any(e['status']!='PASS' or e['returncode']!=0 for e in entries):raise ValueError('Exactly four PASS terminals')
|
||||
if ledger['entries']!=entries or len(models['models'])!=4 or {m['fold'] for m in models['models']}!={1,2,3,4}:raise ValueError('Complete terminal/model manifest')
|
||||
if models['fold0_reused_by_hash']!=tc['fold0_reuse'] or models['new_fits']!=4 or models['test_targets_read'] is not False or models['postfit_started'] is not False:raise ValueError('Model manifest roles')
|
||||
costs=[e['gpu_seconds_charged'] for e in entries]
|
||||
if not all(math.isfinite(x) and x>=0 for x in costs):raise ValueError('Invalid settled GPU cost')
|
||||
total=sum(costs)
|
||||
if done['prior_gpu_once']!=tc['prior_gpu_seconds'] or ledger['prior_gpu_once']!=tc['prior_gpu_seconds'] or ledger['settled_gpu_seconds']!=total or done['new_gpu_seconds_charged']!=total or done['cumulative_h2_gpu_seconds']!=tc['prior_gpu_seconds']+total:raise ValueError('Prior and four new costs must be counted once')
|
||||
if total>tc['train_gpu_cap'] or done['cumulative_h2_gpu_seconds']+2400>144000:raise ValueError('No original prediction reserve')
|
||||
return done['cumulative_h2_gpu_seconds']
|
||||
|
||||
def terminal_inputs(c,pilot,tc):
|
||||
import psutil
|
||||
train=Path(c['training_root']);terminal=verify(train,'terminal');modelproof=verify(train,'models')
|
||||
if (train/'global-stop.json').exists():raise PermissionError('Training stopped')
|
||||
done=read(train/'complete.json');ledger=read(train/'terminal-ledger.json');manifest=read(train/'models-terminal.json');prior=terminal_identity(tc,done,ledger,manifest)
|
||||
for p in psutil.process_iter(['pid','cmdline']):
|
||||
args=p.info['cmdline'] or []
|
||||
if str(train) in args and any(Path(x).name in ['phase2_h2_s2_queue.py','phase2_h2_s2_guard.py','phase2_h2_s2_worker.py'] for x in args):raise RuntimeError('Training process still alive')
|
||||
entries={e['fold']:e for e in done['terminal']};out=[]
|
||||
for m in sorted(manifest['models'],key=lambda m:m['fold']):
|
||||
f=m['fold'];folder=train/'folds'/str(f);guard=read(checked(folder/'complete.json',entries[f]['complete_sha256']));resultpath=checked(folder/'worker/result.json',m['worker_result_sha256']);result=read(resultpath);p=pilot['folds'][f]
|
||||
if guard['status']!='PASS' or guard['reasons'] or guard['monitor_errors'] or guard['worker_result_sha256']!=m['worker_result_sha256']:raise ValueError('Original guard/result identity')
|
||||
if result['fold']!=f or result['completed_epochs']!=200 or result['optimizer_steps']!=200*len(p['learning']) or result['learning_nodes_per_epoch']!=p['learning_nodes'] or result['test_targets_read'] is not False or result['validation_used_for_selection'] is not False:raise ValueError('Final200/role identity')
|
||||
expected=folder/'worker/last-epoch200.pt';stdpath=folder/'worker/standardization.json'
|
||||
if Path(m['checkpoint']).resolve()!=expected.resolve() or Path(m['standardization']).resolve()!=stdpath.resolve() or m['checkpoint_sha256']!=result['checkpoint_sha256'] or m['standardization_sha256']!=result['standardization_sha256']:raise ValueError('Checkpoint/normalization path identity')
|
||||
for path,digest in [(expected,m['checkpoint_sha256']),(stdpath,m['standardization_sha256'])]:
|
||||
checked(path,digest)
|
||||
if modelproof['files'].get(str(path.resolve()))!=digest:raise ValueError('Checkpoint/normalization omitted from model freeze')
|
||||
stats=read(stdpath)
|
||||
if stats['parts']!=p['learning'] or stats['nodes']!=p['learning_nodes'] or len(stats['mean'])!=19 or len(stats['scale'])!=19 or not np.isfinite(stats['mean']).all() or not np.isfinite(stats['scale']).all() or min(stats['scale'])<=0:raise ValueError('Original learning-only normalization')
|
||||
out.append(m)
|
||||
return dict(training_terminal_commit=terminal['commit'],training_models_commit=modelproof['commit'],training_complete_sha256=sha(train/'complete.json'),training_models_sha256=sha(train/'models-terminal.json'),training_contract_sha256=c['source_sha256'][str(train/'contract.json')],models=out,prior_gpu_seconds=prior,old_s1_gpu_once=tc['prior_gpu_seconds'],new_training_gpu_once=done['new_gpu_seconds_charged'])
|
||||
|
||||
def adopted_inputs(root,c,pilot,tc):
|
||||
verify(root,'models');a=read(root/'models-adopted.json')
|
||||
if a!=terminal_inputs(c,pilot,tc):raise ValueError('Adopted four frozen checkpoints changed')
|
||||
return a
|
||||
|
||||
def checkpoint_identity(blob,c,fold,parts):
|
||||
if fold not in (1,2,3,4) or blob['completed_epochs']!=200 or blob['optimizer_steps']!=200*parts or blob['seed']!=0 or blob['fold']!=fold or blob['role']!='last_epoch_only' or blob['contract_sha256']!=c['source_sha256'][str(Path(c['training_root'])/'contract.json')]:raise ValueError('Wrong fold/final checkpoint')
|
||||
|
||||
def comparisons(rows):
|
||||
out={}
|
||||
for reference in ['A','B13_E1b']:
|
||||
raw=pooled([dict(component_id=r['component_id'],new=r['H2'],A=r[reference]) for r in rows]);out[reference]={}
|
||||
for name,m in raw.items():out[reference][name]=dict(means={'H2':m['means']['new'],reference:m['means']['A']},components={'H2':m['components']['new'],reference:m['components']['A']},H2_minus_reference=m['new_minus_A'])
|
||||
return out
|
||||
|
||||
def reuse_fold0_metric(prior,baseline,entry):
|
||||
if entry['fold']!=0 or prior['fold']!=0 or prior['component_id']!=entry['component_id'] or prior['A']!=baseline['A']:raise ValueError('Fold0 frozen score identity')
|
||||
return prior['new']
|
||||
|
||||
def decision(results,parts,groups):
|
||||
field=results['A']['relative_l2'];paired=field['H2_minus_reference']
|
||||
if parts!=552 or groups!=204 or field['components']!={'H2':204,'A':204} or paired['components']!=204:return 'PARK_INCOMPLETE_FIELD_SUPPORT'
|
||||
return 'CONFIRMED' if confirmed(paired['high'],parts,groups) else 'NOT_CONFIRMED_FIXED_RECIPE'
|
||||
|
||||
def require_budget(prior,gpu_spent,cpu_spent,remaining_predictions):
|
||||
if not all(math.isfinite(x) and x>=0 for x in [prior,gpu_spent,cpu_spent]) or remaining_predictions not in range(5):raise ValueError('Invalid resource ledger')
|
||||
if gpu_spent+remaining_predictions*600>2400 or prior+gpu_spent+remaining_predictions*600>144000:raise RuntimeError('Prediction GPU reservation')
|
||||
# All four60-second local CPU setup allowances are already in setup_cpu_seconds=360.
|
||||
if cpu_spent+remaining_predictions*540+600+120>3600:raise RuntimeError('Postfit CPU reservation')
|
||||
@@ -0,0 +1,74 @@
|
||||
"""S2 one-fold600 GPU-second prediction, retained failure peaks, no retries."""
|
||||
import json,os,signal,sys,threading,time
|
||||
import psutil
|
||||
from phase2_h1_worker_v2 import save
|
||||
from s2_main_memory_guard import PeakRecords
|
||||
from s2_precision_continuation_guard import device_usage
|
||||
from s2_recovery_gpu_supervisor import supervise,tree_usage,stop_group,disk_usage
|
||||
|
||||
def final_reasons(reasons,receipt,result,tracker,gpu_seconds,cpu_seconds,cpu_cap,expected_parts):
|
||||
out=list(reasons)
|
||||
def add(s):
|
||||
if s not in out:out.append(s)
|
||||
if not receipt or receipt['status']!='PASS':add('WORKER_NOT_PASS')
|
||||
if not result or result.get('status')!='COMPLETE' or result.get('parts')!=expected_parts:add('INCOMPLETE_FOLD')
|
||||
if not all(k in tracker.witnesses for k in ['peak_allocated_bytes','peak_reserved_bytes']):add('MISSING_PEAKS')
|
||||
if any(v>15*2**30 for v in tracker.maxima.values()):add('RECORDED_VRAM_LIMIT')
|
||||
if gpu_seconds>600 or cpu_seconds>cpu_cap:add('FINAL_RESOURCE_LIMIT')
|
||||
return out
|
||||
|
||||
def run_prediction(queue,c,pilot,fold,prior_gpu,cpu_cap):
|
||||
root=queue/'folds'/str(fold);gpu_uuid=c['training_jobs'][fold-1]['gpu_uuid'];expected_parts=len(pilot['folds'][fold]['test'])
|
||||
start=time.monotonic();cpu=time.process_time();requested=[];reasons=[];errors=[];samples=[];done=threading.Event();tracker=PeakRecords();receipt=None
|
||||
def elapsed():return 60+time.monotonic()-start
|
||||
def reason(x):
|
||||
if x not in reasons:reasons.append(x)
|
||||
command=[sys.executable,'tools/phase2_h2_s2_postfit.py','predict',str(queue),str(fold)]
|
||||
def kill():
|
||||
for p in psutil.Process().children(recursive=True):
|
||||
try:
|
||||
if p.cmdline()==command:stop_group(p.pid)
|
||||
except psutil.NoSuchProcess:pass
|
||||
def scan():
|
||||
if tracker.scan(root/'predictions'):reason('RECORDED_VRAM_LIMIT')
|
||||
def watch():
|
||||
while not done.is_set():
|
||||
try:
|
||||
scan();mem,util=device_usage(gpu_uuid);_,rss,_=tree_usage(os.getpid())
|
||||
if mem>15*2**30:reason('DEVICE_VRAM_LIMIT')
|
||||
if rss>16*2**30:reason('TREE_RSS_LIMIT')
|
||||
if disk_usage(queue)>20*2**30:reason('DISK_LIMIT')
|
||||
if elapsed()>=595:reason('PREDICTION_TIME_LIMIT')
|
||||
for x in requested:reason(x)
|
||||
save(root/'prediction-live.json',dict(gpu_seconds_charged=elapsed(),device_bytes=mem,device_utilization=util,rss_bytes=rss,peaks=tracker.maxima,reasons=reasons,unix=time.time()))
|
||||
except Exception as exc:reason('MONITOR_ERROR');errors.append(str(exc))
|
||||
if reasons:kill()
|
||||
done.wait(.5)
|
||||
old={s:signal.signal(s,lambda signum,frame:requested.append('COORDINATOR_SIGNAL_'+str(signum))) for s in [signal.SIGTERM,signal.SIGINT]};thread=None
|
||||
try:
|
||||
for i in range(3):
|
||||
mem,util=device_usage(gpu_uuid);samples.append(dict(unix=time.time(),bytes=mem,utilization=util))
|
||||
if mem>128*2**20 or util!=0:raise RuntimeError('SELECTED_GPU_BUSY_NO_RETRY')
|
||||
if i<2:time.sleep(2)
|
||||
save(root/'prediction-idle-polls.json',samples)
|
||||
if requested:raise RuntimeError('Signal received during setup')
|
||||
thread=threading.Thread(target=watch,daemon=True);thread.start()
|
||||
remaining=600-elapsed()-10
|
||||
if remaining<=0:raise RuntimeError('Setup exhausted prediction budget')
|
||||
receipt=supervise(command,root/'prediction-guard','predict',dict(total_cpu_seconds=cpu_cap-60,job_cpu_seconds=cpu_cap-60,job_wall_seconds=remaining,tree_rss_bytes=16*2**30,disk_bytes=20*2**30,max_processes=1),poll=.2,gpu_uuid=gpu_uuid)
|
||||
except BaseException as exc:reason(type(exc).__name__+': '+str(exc));kill()
|
||||
finally:
|
||||
done.set()
|
||||
if thread:thread.join(timeout=4)
|
||||
if thread and thread.is_alive():reason('MONITOR_DID_NOT_EXIT')
|
||||
try:scan()
|
||||
except Exception as exc:reason('FINAL_SCAN_ERROR');errors.append(str(exc))
|
||||
for s,h in old.items():signal.signal(s,h)
|
||||
result=None;p=root/'predictions/result.json'
|
||||
try:
|
||||
if p.exists():result=json.loads(p.read_text())
|
||||
except Exception as exc:reason('FINAL_RESULT_READ_ERROR');errors.append(str(exc))
|
||||
totalcpu=60+(receipt['cpu_seconds'] if receipt else 0)+time.process_time()-cpu
|
||||
duration=elapsed();reasons=final_reasons(reasons,receipt,result,tracker,duration,totalcpu,cpu_cap,expected_parts)
|
||||
out=dict(status='PASS' if not reasons else 'STOPPED',reasons=reasons,monitor_errors=errors,worker=receipt,recorded_memory=tracker.as_dict(),gpu_seconds_charged=duration,cpu_seconds_charged=totalcpu,cumulative_h2_gpu_seconds=prior_gpu+duration,no_retry=True)
|
||||
save(root/'prediction-guard.json',out);return out
|
||||
@@ -0,0 +1,62 @@
|
||||
"""Prepare exact S2 review binding while original training continues; no launch."""
|
||||
import json,shutil,time
|
||||
from pathlib import Path
|
||||
from phase2_fixed_io import BASE,verify as old_verify
|
||||
from phase2_home_freeze_io import freeze,verify
|
||||
from phase2_h1_worker_v2 import checked,sha,save
|
||||
from phase2_h2_s2_common import config as training_config
|
||||
from phase2_h2_s2_postfit_common import read,roster,config
|
||||
|
||||
def main():
|
||||
root=Path('/home/frank14f/MeltMind-runs/phase2-h2-s2-postfit-2026-11-17')
|
||||
if root.exists():raise FileExistsError('Existing S2 postfit; inspect, never duplicate')
|
||||
train=Path('/home/frank14f/MeltMind-runs/phase2-h2-s2-2026-11-17');tc,pilot=training_config(train);tp=verify(train,'implementation')
|
||||
if tp['commit']!='1f57bad66eb062d4f2269347532321929f873717' or sha(train/'contract.json')!='76e1976bed88fc390812238c3d80693446daea3840dc009ac7d8870ddc43a31d':raise ValueError('Original pinned training identity')
|
||||
s1=Path(tc['s1_root']);sc=read(s1/'contract.json');s1pred=verify(s1,'predictions');s1terminal=verify(s1,'terminal')
|
||||
aroot=BASE/'artifacts/phase2-b13-e1b-2026-11-17';a=read(aroot/'contract.json');old_verify(aroot,'predictions');old_verify(aroot,'terminal-result')
|
||||
hroot=BASE/'artifacts/phase2-h1-execution-2026-11-17';h=read(hroot/'contract.json');jobs={j['fold']:j for j in h['jobs'] if j['protocol']=='grouped' and j['pool']==100}
|
||||
rows=[];models=[]
|
||||
for f in pilot['folds']:
|
||||
af=a['folds'][f['fold']];hj=jobs[f['fold']]
|
||||
if hj['baseline_model']!=af['A_baseline'] or set(hj['learning'])!=set(f['learning']) or set(hj['test'])!=set(f['test']):raise ValueError('H1/A/H2 original roles or baseline model differ')
|
||||
if {r['case'] for r in af['rows']['learning']}!=set(f['learning']) or {r['case'] for r in af['rows']['test']}!=set(f['test']):raise ValueError('Original E1b learning/test roles differ')
|
||||
by={r['case']:r for r in af['rows']['test']};rows.extend(by[k] for k in f['test']);models.append(hj['baseline_model'])
|
||||
entries=read(aroot/'predictions-manifest.json')['entries'];baseline={e['case']:e for e in entries};fold0=read(s1/'predictions-manifest.json')['entries']
|
||||
if len(entries)!=552 or len(baseline)!=552 or len(fold0)!=121 or {e['case'] for e in fold0}!=set(pilot['folds'][0]['test']):raise ValueError('Frozen prediction manifest coverage')
|
||||
for e in entries:
|
||||
if e['A_source_sha256']!=a['A_predictions'][e['case']]['sha256']:raise ValueError('Frozen A source identity')
|
||||
for e in fold0:
|
||||
if e['fold']!=0 or e['A_source_sha256']!=baseline[e['case']]['sha256']:raise ValueError('S1 original baseline pairing')
|
||||
sources={}
|
||||
def add(p,digest):
|
||||
p=str((BASE/Path(p)).resolve())
|
||||
if p in sources and sources[p]!=digest:raise ValueError('Conflicting input hash: '+p)
|
||||
sources[p]=digest
|
||||
for old in [tc,sc]:
|
||||
for p,digest in old['source_sha256'].items():add(p,digest)
|
||||
for p in [train/'contract.json',train/'implementation-git-freeze.json',Path(tc['pilot_contract']),s1/'contract.json',s1/'predictions-manifest.json',s1/'predictions-git-freeze.json',s1/'terminal-git-freeze.json',s1/'complete.json',s1/'post-terminal-cost.json',s1/'score/result.json',aroot/'contract.json',aroot/'predictions-manifest.json',aroot/'score/result.json',aroot/'predictions-git-freeze.json',aroot/'terminal-result-git-freeze.json',hroot/'contract.json']:add(p,sha(p))
|
||||
for e in entries+fold0:add(e['path'],e['sha256'])
|
||||
for m in models:add(m['path'],m['sha256'])
|
||||
for p,digest in sources.items():checked(p,digest) # Opaque bytes only; no score/target value parsing.
|
||||
codes={}
|
||||
for old in [tc,sc,a]:
|
||||
for p,digest in old['code_sha256'].items():
|
||||
p=str((BASE/Path(p)).resolve())
|
||||
if p in codes and codes[p]!=digest:raise ValueError('Conflicting source pin: '+p)
|
||||
checked(p,digest);codes[p]=digest
|
||||
for name in ['phase2_h2_s2_postfit_common.py','phase2_h2_s2_postfit.py','phase2_h2_s2_prediction_guard.py','prepare_phase2_h2_s2_postfit.py','test_phase2_h2_s2_postfit.py']:
|
||||
p=BASE/'tools'/name;codes[str(p)]=sha(p)
|
||||
doc=BASE/'docs/phase2-h2-s2-postfit-2026-11-17.md';codes[str(doc)]=sha(doc)
|
||||
comparison=dict(approval='1794925470.101149',test_parts=552,test_components=204,fold0_prediction_and_metric_hash_reuse=True,new_prediction_parts=431,checkpoint_folds=[1,2,3,4],checkpoint_epoch=200,seed=0,A_models=models,clocks=[dict(fold=f['fold'],c=f['c'],path=f['clock_path'],sha256=f['clock_sha256']) for f in pilot['folds']],metrics=['relative_l2','last_location','kendall_tau_b'],query='Original eligible S2 queries; multiplicities once; raw-ID minimum tie rule; last distance/Rg',pooling='Original E1b pooled(): per-arm component requires all constituent part metrics defined; component means equally weighted; paired intersection',bootstrap=dict(replicates=2000,seed=2026111408),confirmation='CONFIRMED iff complete552/204 paired field H2-A upper95CI<0; strict finite bound; incomplete field support PARK',report_only=['H2-B13_E1b','last_location','kendall_tau_b'],negative_clipping=False,all_four_checkpoints_frozen_and_training_exited_before_prediction=True,actual_new_semantic_PASS_before_any_target=True,predictions_frozen_before_score=True,new_fits=0,new_decodes=0,no_retry=True,no_follow_on_launch=True,interpretation='Five-fold recipe confirmation after observing fold0; not an untouched-five-fold preregistration or connectivity causal claim')
|
||||
log=Path('/tmp/h2-s2-postfit-synthetic-attempt1.log').read_text()
|
||||
if 'Ran 23 tests' not in log or not log.rstrip().endswith('OK'):raise ValueError('Actual23 synthetic PASS required')
|
||||
c=dict(version='H2-S2-pooled-postfit-v1',approval='1794925470.101149',postfit_root=str(root),training_root=str(train),training_implementation=tp['commit'],training_jobs=tc['jobs'],s1_root=str(s1),pilot_contract=tc['pilot_contract'],fold0_manifest=str(s1/'predictions-manifest.json'),fold0_score=str(s1/'score/result.json'),baseline_score=str(aroot/'score/result.json'),rows=rows,baseline_predictions=baseline,source_sha256=sources,code_sha256=codes,total_h2_gpu_seconds=144000,prediction_gpu_seconds=2400,postfit_cpu_seconds=3600,setup_cpu_seconds=360,prediction_per_fold_gpu_seconds=600,prediction_per_fold_worker_cpu_seconds=540,score_worker_cpu_seconds=600,terminal_cpu_reserve=120,coordinator_cpus=[12,13],vram_bytes=15*2**30,rss_bytes=16*2**30,disk_bytes=20*2**30,comparison_contract=str(root/'comparison-contract.json'),new_fits=0,new_decodes=0,no_retry=True,actual_new_semantic_PASS_required=True,all_four_terminal_before_prediction=True,no_auto_launch=True)
|
||||
roster(c,pilot);root.mkdir();save(root/'comparison-contract.json',comparison);c['comparison_contract_sha256']=sha(root/'comparison-contract.json');save(root/'contract.json',c)
|
||||
shutil.copyfile('/tmp/h2-s2-postfit-synthetic-attempt1.log',root/'synthetic-attempt1.log')
|
||||
save(root/'acceptance.json',dict(status='PASS',tests=23,elapsed_seconds=1.477,log_sha256=sha(root/'synthetic-attempt1.log'),test_source_sha256=sha(BASE/'tools/test_phase2_h2_s2_postfit.py'),scope='CPU synthetic plus actual isolated Git/file review-validator tests; no real prediction, target read, fitting or CUDA workload',setup_cpu_seconds_reserved=360,unix=time.time()))
|
||||
save(root/'preparation-audit.json',dict(status='PREPARED_NOT_LAUNCHED',unix=time.time(),training_implementation=tp['commit'],S1_prediction_commit=s1pred['commit'],S1_terminal_commit=s1terminal['commit'],test_parts=552,test_components=204,fold_test_counts=[len(f['test']) for f in pilot['folds']],original_learning_test_roles_and_A_models_equal=True,original_prediction_hashes_checked=673,new_fits=0,new_predictions=0,new_scores=0,old_score_values_read=False,test_target_values_read=False,training_untouched=True,gate_queue_untouched=True,future_checkpoint_adoption='Not present or required at preparation; all four original models/terminal freezes and process exit mandatory at execution',actual_new_review_pending=True))
|
||||
freeze(root,'implementation',[root/'contract.json',root/'comparison-contract.json',root/'acceptance.json',root/'synthetic-attempt1.log',root/'preparation-audit.json',*[Path(p) for p in codes]])
|
||||
proof=verify(root,'implementation');save(root/'expected-review-binding.json',dict(reviewed_commit=proof['commit'],contract_sha256=sha(root/'contract.json'),implementation_proof_sha256=sha(root/'implementation-git-freeze.json'),comparison_contract_sha256=c['comparison_contract_sha256'],code_sha256=codes))
|
||||
freeze(root,'review-binding',[root/'contract.json',root/'expected-review-binding.json',root/'implementation-git-freeze.json',root/'implementation-authorization.json']);config(root)
|
||||
print(json.dumps(dict(root=str(root),implementation=proof['commit'],contract_sha256=sha(root/'contract.json'),binding_sha256=sha(root/'expected-review-binding.json'),code_files=len(codes),review_pending=True,launched=False)))
|
||||
if __name__=='__main__':main()
|
||||
@@ -0,0 +1,101 @@
|
||||
"""Synthetic pooled postfit boundaries; no real model prediction or target values."""
|
||||
import ast,copy,inspect,json,tempfile,unittest
|
||||
from pathlib import Path
|
||||
from phase2_h2_s2_postfit_common import terminal_identity,terminal_inputs,checkpoint_identity,comparisons,decision,require_budget,reuse_fold0_metric
|
||||
from phase2_h2_s2_postfit import predict,score,manifest_identity
|
||||
from phase2_h2_s2_prediction_guard import final_reasons
|
||||
from s2_main_memory_guard import PeakRecords
|
||||
from test_phase2_gate_outcome_postfit import SemanticReview
|
||||
|
||||
def metric(v):return {k:dict(value=v) for k in ['relative_l2','last_location','kendall_tau_b']}
|
||||
class TerminalAndBudget(unittest.TestCase):
|
||||
def fixture(self):
|
||||
tc=dict(prior_gpu_seconds=18858.21490871499,train_gpu_cap=122741.78509128501,fold0_reuse={'hash':'original'});entries=[dict(fold=f,status='PASS',returncode=0,gpu_seconds_charged=20000.) for f in range(1,5)]
|
||||
done=dict(status='TRAIN_COMPLETE_POSTFIT_PENDING',postfit_started=False,terminal=entries,prior_gpu_once=tc['prior_gpu_seconds'],new_gpu_seconds_charged=80000.,cumulative_h2_gpu_seconds=tc['prior_gpu_seconds']+80000.)
|
||||
ledger=dict(entries=entries,prior_gpu_once=tc['prior_gpu_seconds'],settled_gpu_seconds=80000.);models=dict(models=[dict(fold=f) for f in range(1,5)],fold0_reused_by_hash=tc['fold0_reuse'],new_fits=4,test_targets_read=False,postfit_started=False);return tc,done,ledger,models
|
||||
def test_all_four_costs_and_prior_once(self):
|
||||
a=self.fixture();self.assertEqual(terminal_identity(*a),98858.21490871499)
|
||||
def test_partial_failed_or_wrong_role_models_rejected(self):
|
||||
for kind in ['partial','failed','duplicate','predicted']:
|
||||
t,d,l,m=self.fixture()
|
||||
if kind=='partial':d['terminal']=d['terminal'][:3]
|
||||
if kind=='failed':d['terminal'][0]['status']='FAILED'
|
||||
if kind=='duplicate':m['models'][3]['fold']=1
|
||||
if kind=='predicted':m['test_targets_read']=True
|
||||
with self.assertRaises((ValueError,PermissionError)):terminal_identity(t,d,l,m)
|
||||
def test_mismatched_or_double_counted_ledger_rejected(self):
|
||||
for key in ['prior_gpu_once','new_gpu_seconds_charged','cumulative_h2_gpu_seconds']:
|
||||
t,d,l,m=self.fixture();d[key]+=1
|
||||
with self.assertRaises(ValueError):terminal_identity(t,d,l,m)
|
||||
def test_still_training_without_model_freeze_cannot_enter(self):
|
||||
with tempfile.TemporaryDirectory() as p:
|
||||
with self.assertRaises(FileNotFoundError):terminal_inputs({'training_root':p},{},{})
|
||||
def test_correct_new_fold_final_checkpoint_only(self):
|
||||
c=dict(training_root='/train',source_sha256={'/train/contract.json':'training-fixed'});b=dict(completed_epochs=200,optimizer_steps=73200,seed=0,fold=2,role='last_epoch_only',contract_sha256='training-fixed');checkpoint_identity(b,c,2,366)
|
||||
for k,v in dict(fold=0,completed_epochs=199,optimizer_steps=64000,contract_sha256='s1',role='best').items():
|
||||
with self.assertRaises(ValueError):checkpoint_identity(dict(b,**{k:v}),c,2,366)
|
||||
def test_budget_reserves_all_remaining_predictions_and_score(self):
|
||||
require_budget(98858.,0,360,4);require_budget(98858.,300,800,1)
|
||||
for args in [(142000,0,360,4),(98858,2200,500,1),(98858,0,1000,4),(98858,0,3000,0)]:
|
||||
with self.assertRaises(RuntimeError):require_budget(*args)
|
||||
def test_prediction_final_peaks_and_per_fold_coverage(self):
|
||||
t=PeakRecords();t.witnesses={'peak_allocated_bytes':{},'peak_reserved_bytes':{}}
|
||||
self.assertEqual(final_reasons([],{'status':'PASS'},dict(status='COMPLETE',parts=91),t,100,100,600,91),[])
|
||||
self.assertIn('INCOMPLETE_FOLD',final_reasons([],{'status':'PASS'},dict(status='COMPLETE',parts=121),t,100,100,600,91))
|
||||
t.maxima['peak_reserved_bytes']=16*2**30
|
||||
self.assertIn('RECORDED_VRAM_LIMIT',final_reasons([],{'status':'FAILED'},None,t,100,100,600,91))
|
||||
self.assertIn('FINAL_RESOURCE_LIMIT',final_reasons([],{'status':'PASS'},dict(status='COMPLETE',parts=91),t,601,100,600,91))
|
||||
|
||||
class StatisticsAndCustody(unittest.TestCase):
|
||||
def test_component_equal_weights_and_b13_report_comparison(self):
|
||||
rows=[dict(component_id=g,H2=metric(h),A=metric(a),B13_E1b=metric(b)) for g,h,a,b in [('x',2,1,0),('x',2,1,0),('y',1,1,0)]];d=comparisons(rows)
|
||||
self.assertEqual(d['A']['relative_l2']['H2_minus_reference']['value'],.5);self.assertEqual(d['B13_E1b']['relative_l2']['H2_minus_reference']['value'],1.5)
|
||||
for arm in d:
|
||||
self.assertEqual(d[arm]['relative_l2']['H2_minus_reference']['replicates'],2000);self.assertEqual(d[arm]['relative_l2']['H2_minus_reference']['seed'],2026111408)
|
||||
def test_original_whole_component_undefined_rule(self):
|
||||
rows=[dict(component_id=g,H2=metric(h),A=metric(1),B13_E1b=metric(1)) for g,h in [('x',2),('x',None),('y',1)]];d=comparisons(rows);self.assertEqual(d['A']['relative_l2']['components'],{'H2':1,'A':2});self.assertEqual(d['A']['relative_l2']['H2_minus_reference']['components'],1)
|
||||
def test_only_strict_complete_field_vs_A_confirms(self):
|
||||
def result(high):return {'A':{'relative_l2':dict(components={'H2':204,'A':204},H2_minus_reference=dict(components=204,high=high))},'B13_E1b':{'relative_l2':{'arbitrarily_bad_report_only':999}}}
|
||||
self.assertEqual(decision(result(-.001),552,204),'CONFIRMED')
|
||||
for high in [0,.001,None,float('nan')]:self.assertEqual(decision(result(high),552,204),'NOT_CONFIRMED_FIXED_RECIPE')
|
||||
self.assertEqual(decision(result(-.001),551,204),'PARK_INCOMPLETE_FIELD_SUPPORT')
|
||||
r=result(-.001);r['A']['relative_l2']['components']['H2']=203;self.assertEqual(decision(r,552,204),'PARK_INCOMPLETE_FIELD_SUPPORT')
|
||||
def test_fold0_metrics_are_reused_not_recomputed(self):
|
||||
h=metric(.25);a=metric(.5);prior=dict(fold=0,component_id='x',new=h,A=a);entry=dict(fold=0,component_id='x');self.assertIs(reuse_fold0_metric(prior,{'A':a},entry),h)
|
||||
with self.assertRaises(ValueError):reuse_fold0_metric(prior,{'A':metric(.6)},entry)
|
||||
def test_fold0_prediction_forbidden_before_config_or_CUDA(self):
|
||||
with self.assertRaises(PermissionError):predict(Path('/nonexistent'),0)
|
||||
def test_semantic_pass_and_all_models_precede_value_reads(self):
|
||||
s=inspect.getsource(score)
|
||||
for token in ["c['baseline_score']","c['fold0_score']","cached(row,'seconds')"]:
|
||||
self.assertLess(s.index('validate_review(root)'),s.index(token));self.assertLess(s.index('adopted_inputs('),s.index(token))
|
||||
node=ast.parse(s);branch=next(n for n in ast.walk(node) if isinstance(n,ast.If) and ast.unparse(n.test)=="e['fold'] == 0")
|
||||
self.assertNotIn('cached(',ast.unparse(branch.body));self.assertIn('reuse_fold0_metric',ast.unparse(branch.body));self.assertIn("cached(row, 'seconds')",ast.unparse(branch.orelse))
|
||||
|
||||
class ManifestReuse(unittest.TestCase):
|
||||
def setUp(self):
|
||||
self.tmp=tempfile.TemporaryDirectory();self.root=Path(self.tmp.name);self.rows=[];parts={};baseline={};folds=[dict(fold=f,test=[]) for f in range(5)];self.entries=[];old=[]
|
||||
for i in range(552):
|
||||
case='case'+str(i);g='g'+str(i%204);f=(i%204)%5;cache={'seconds.npy':'label-digest'};folds[f]['test'].append(case);parts[case]=dict(nodes=3,clock_group=g,output_sha256=cache,t_set=.4);baseline[case]=dict(fold=f,component_id=g,sha256='baseline-'+case)
|
||||
self.rows.append(dict(case=case,component_id=g,nodes=3,assignments=[dict(fold=f,role='test')],v5_cache={'output_sha256':cache},preparation={'t_set':.4}))
|
||||
e=dict(case=case,component_id=g,fold=f,nodes=3,query_nodes=2,path=str(self.root/'folds'/str(f)/'predictions'/(case+'.npz')),sha256='prediction-'+case,A_source_sha256='baseline-'+case,negative_full_predictions=1,negative_query_predictions=1)
|
||||
if f==0:old.append(dict(e))
|
||||
self.entries.append(dict(e,origin='S1_HASH_REUSE' if f==0 else 'S2_FROZEN_CHECKPOINT'))
|
||||
p=self.root/'old.json';p.write_text(json.dumps(dict(entries=old)));self.c=dict(rows=self.rows,baseline_predictions=baseline,fold0_manifest=str(p),postfit_root=str(self.root));self.pilot=dict(folds=folds,parts=parts)
|
||||
def tearDown(self):self.tmp.cleanup()
|
||||
def test_complete_roster_and_exact_fold0_entries(self):self.assertEqual(len(manifest_identity(self.entries,self.c,self.pilot)),552)
|
||||
def test_fold0_changed_hash_rejected(self):
|
||||
d=copy.deepcopy(self.entries);next(e for e in d if e['fold']==0)['sha256']='different'
|
||||
with self.assertRaises(ValueError):manifest_identity(d,self.c,self.pilot)
|
||||
def test_duplicate_missing_and_wrong_new_fold_path_rejected(self):
|
||||
for kind in ['duplicate','missing','path']:
|
||||
d=copy.deepcopy(self.entries)
|
||||
if kind=='duplicate':d[-1]=d[0]
|
||||
if kind=='missing':d.pop()
|
||||
if kind=='path':next(e for e in d if e['fold']==1)['path']='/other/pred.npz'
|
||||
with self.assertRaises(ValueError):manifest_identity(d,self.c,self.pilot)
|
||||
def test_negative_counts_not_silently_changed_or_boolean(self):
|
||||
for n in [-1,4,True]:
|
||||
d=copy.deepcopy(self.entries);next(e for e in d if e['fold']==1)['negative_full_predictions']=n
|
||||
with self.assertRaises(ValueError):manifest_identity(d,self.c,self.pilot)
|
||||
if __name__=='__main__':unittest.main()
|
||||
Reference in new issue
Block a user