summaryrefslogtreecommitdiff
path: root/runs/pe3_sweep.log
diff options
context:
space:
mode:
Diffstat (limited to 'runs/pe3_sweep.log')
-rw-r--r--runs/pe3_sweep.log220
1 files changed, 220 insertions, 0 deletions
diff --git a/runs/pe3_sweep.log b/runs/pe3_sweep.log
new file mode 100644
index 0000000..8092566
--- /dev/null
+++ b/runs/pe3_sweep.log
@@ -0,0 +1,220 @@
+host=timan1.cs.illinois.edu gpu=0 start=2026-06-16T13:42:05-05:00
+===== seed=0 #3 full sub =====
+/home/yurenh2/rrog/diag/train_color.py:86: RuntimeWarning: invalid value encountered in divide
+ clus = np.where(deg > 1, tri / (deg * (deg - 1) / 2.0), 0.0)
+ep20 solve_rate=0.000 mean_conflicts=56.87
+ep40 solve_rate=0.280 mean_conflicts=13.94
+ep60 solve_rate=0.207 mean_conflicts=44.20
+ep80 solve_rate=0.440 mean_conflicts=39.18
+ep100 solve_rate=0.710 mean_conflicts=1.94
+ep120 solve_rate=0.643 mean_conflicts=7.74
+ep140 solve_rate=0.530 mean_conflicts=26.79
+ep150 solve_rate=0.500 mean_conflicts=28.63
+[color_full_sub_n50_k3_p0.2_T3_ns3_s0] best solve_rate=0.71 mean_conflicts=1.94 @ep100 (155.3s)
+ wrote /home/yurenh2/rrog/runs/ckpt_color_full_sub_n50_k3_p0.2_T3_ns3_s0.pt
+/home/yurenh2/rrog/diag/train_color.py:86: RuntimeWarning: invalid value encountered in divide
+ clus = np.where(deg > 1, tri / (deg * (deg - 1) / 2.0), 0.0)
+[full] LE n=300 fail_rate=0.29 | lambda1 SOLVED mean -0.3371 (n=213) | UNSOLVED mean +0.0344 (n=87) | sep=+0.3714 | AUROC(fail|lambda1)=0.885 | mean_lambda1=-0.2293
+/home/yurenh2/rrog/diag/train_color.py:86: RuntimeWarning: invalid value encountered in divide
+ clus = np.where(deg > 1, tri / (deg * (deg - 1) / 2.0), 0.0)
+[pe=sub s0] deterministic solve_rate = 0.733 (n=150, K=16)
+ sigma pass@K lam-sel random perRoll AUROC(s|-lam)
+ 0.2 0.920 0.860 0.727 0.732 0.912
+ wrote /home/yurenh2/rrog/runs/ptrm_color_full_sub_n50_k3_p0.2_T3_ns3_s0.json
+===== seed=0 #4 full none --contract =====
+ep20 solve_rate=0.000 mean_conflicts=163.48
+ep40 solve_rate=0.000 mean_conflicts=146.81
+ep60 solve_rate=0.000 mean_conflicts=135.64
+ep80 solve_rate=0.000 mean_conflicts=126.22
+ep100 solve_rate=0.000 mean_conflicts=113.24
+ep120 solve_rate=0.000 mean_conflicts=91.74
+ep140 solve_rate=0.000 mean_conflicts=88.39
+ep150 solve_rate=0.000 mean_conflicts=87.80
+[color_full_none_ctr_n50_k3_p0.2_T3_ns3_s0] best solve_rate=0.0 mean_conflicts=163.477 @ep20 (207.1s)
+ wrote /home/yurenh2/rrog/runs/ckpt_color_full_none_ctr_n50_k3_p0.2_T3_ns3_s0.pt
+[full] LE n=300 fail_rate=1.00 | lambda1 SOLVED mean +nan (n=0) | UNSOLVED mean -0.6308 (n=300) | sep=+nan | AUROC(fail|lambda1)=nan | mean_lambda1=-0.6308
+[pe=none s0] deterministic solve_rate = 0.000 (n=150, K=16)
+ sigma pass@K lam-sel random perRoll AUROC(s|-lam)
+ 0.2 0.000 0.000 0.000 0.000 nan
+ wrote /home/yurenh2/rrog/runs/ptrm_color_full_none_ctr_n50_k3_p0.2_T3_ns3_s0.json
+===== seed=1 #3 full sub =====
+/home/yurenh2/rrog/diag/train_color.py:86: RuntimeWarning: invalid value encountered in divide
+ clus = np.where(deg > 1, tri / (deg * (deg - 1) / 2.0), 0.0)
+ep20 solve_rate=0.000 mean_conflicts=99.67
+ep40 solve_rate=0.210 mean_conflicts=28.08
+ep60 solve_rate=0.233 mean_conflicts=59.81
+ep80 solve_rate=0.570 mean_conflicts=7.48
+ep100 solve_rate=0.303 mean_conflicts=60.79
+ep120 solve_rate=0.187 mean_conflicts=69.15
+ep140 solve_rate=0.107 mean_conflicts=63.47
+ep150 solve_rate=0.107 mean_conflicts=64.31
+[color_full_sub_n50_k3_p0.2_T3_ns3_s1] best solve_rate=0.57 mean_conflicts=7.483 @ep80 (148.6s)
+ wrote /home/yurenh2/rrog/runs/ckpt_color_full_sub_n50_k3_p0.2_T3_ns3_s1.pt
+/home/yurenh2/rrog/diag/train_color.py:86: RuntimeWarning: invalid value encountered in divide
+ clus = np.where(deg > 1, tri / (deg * (deg - 1) / 2.0), 0.0)
+[full] LE n=300 fail_rate=0.43 | lambda1 SOLVED mean -0.3139 (n=171) | UNSOLVED mean +0.0323 (n=129) | sep=+0.3462 | AUROC(fail|lambda1)=0.928 | mean_lambda1=-0.1650
+/home/yurenh2/rrog/diag/train_color.py:86: RuntimeWarning: invalid value encountered in divide
+ clus = np.where(deg > 1, tri / (deg * (deg - 1) / 2.0), 0.0)
+[pe=sub s1] deterministic solve_rate = 0.540 (n=150, K=16)
+ sigma pass@K lam-sel random perRoll AUROC(s|-lam)
+ 0.2 0.867 0.713 0.540 0.570 0.893
+ wrote /home/yurenh2/rrog/runs/ptrm_color_full_sub_n50_k3_p0.2_T3_ns3_s1.json
+===== seed=1 #4 full none --contract =====
+ep20 solve_rate=0.000 mean_conflicts=107.59
+ep40 solve_rate=0.000 mean_conflicts=97.27
+ep60 solve_rate=0.000 mean_conflicts=84.15
+ep80 solve_rate=0.000 mean_conflicts=75.86
+ep100 solve_rate=0.000 mean_conflicts=99.62
+ep120 solve_rate=0.000 mean_conflicts=163.03
+ep140 solve_rate=0.000 mean_conflicts=162.94
+ep150 solve_rate=0.000 mean_conflicts=162.62
+[color_full_none_ctr_n50_k3_p0.2_T3_ns3_s1] best solve_rate=0.0 mean_conflicts=107.593 @ep20 (206.4s)
+ wrote /home/yurenh2/rrog/runs/ckpt_color_full_none_ctr_n50_k3_p0.2_T3_ns3_s1.pt
+[full] LE n=300 fail_rate=1.00 | lambda1 SOLVED mean +nan (n=0) | UNSOLVED mean -1.0479 (n=300) | sep=+nan | AUROC(fail|lambda1)=nan | mean_lambda1=-1.0479
+[pe=none s1] deterministic solve_rate = 0.000 (n=150, K=16)
+ sigma pass@K lam-sel random perRoll AUROC(s|-lam)
+ 0.2 0.000 0.000 0.000 0.000 nan
+ wrote /home/yurenh2/rrog/runs/ptrm_color_full_none_ctr_n50_k3_p0.2_T3_ns3_s1.json
+===== seed=2 #3 full sub =====
+/home/yurenh2/rrog/diag/train_color.py:86: RuntimeWarning: invalid value encountered in divide
+ clus = np.where(deg > 1, tri / (deg * (deg - 1) / 2.0), 0.0)
+ep20 solve_rate=0.000 mean_conflicts=143.77
+ep40 solve_rate=0.280 mean_conflicts=19.26
+ep60 solve_rate=0.520 mean_conflicts=15.04
+ep80 solve_rate=0.670 mean_conflicts=3.32
+ep100 solve_rate=0.620 mean_conflicts=20.58
+ep120 solve_rate=0.403 mean_conflicts=60.92
+ep140 solve_rate=0.390 mean_conflicts=58.32
+ep150 solve_rate=0.377 mean_conflicts=55.34
+[color_full_sub_n50_k3_p0.2_T3_ns3_s2] best solve_rate=0.67 mean_conflicts=3.317 @ep80 (152.1s)
+ wrote /home/yurenh2/rrog/runs/ckpt_color_full_sub_n50_k3_p0.2_T3_ns3_s2.pt
+/home/yurenh2/rrog/diag/train_color.py:86: RuntimeWarning: invalid value encountered in divide
+ clus = np.where(deg > 1, tri / (deg * (deg - 1) / 2.0), 0.0)
+[full] LE n=300 fail_rate=0.33 | lambda1 SOLVED mean -0.3650 (n=201) | UNSOLVED mean +0.0472 (n=99) | sep=+0.4122 | AUROC(fail|lambda1)=0.937 | mean_lambda1=-0.2290
+/home/yurenh2/rrog/diag/train_color.py:86: RuntimeWarning: invalid value encountered in divide
+ clus = np.where(deg > 1, tri / (deg * (deg - 1) / 2.0), 0.0)
+[pe=sub s2] deterministic solve_rate = 0.707 (n=150, K=16)
+ sigma pass@K lam-sel random perRoll AUROC(s|-lam)
+ 0.2 0.893 0.800 0.700 0.688 0.919
+ wrote /home/yurenh2/rrog/runs/ptrm_color_full_sub_n50_k3_p0.2_T3_ns3_s2.json
+===== seed=2 #4 full none --contract =====
+ep20 solve_rate=0.000 mean_conflicts=102.42
+ep40 solve_rate=0.000 mean_conflicts=89.46
+ep60 solve_rate=0.000 mean_conflicts=83.90
+ep80 solve_rate=0.000 mean_conflicts=90.79
+ep100 solve_rate=0.000 mean_conflicts=92.59
+ep120 solve_rate=0.000 mean_conflicts=84.45
+ep140 solve_rate=0.000 mean_conflicts=76.00
+ep150 solve_rate=0.000 mean_conflicts=69.42
+[color_full_none_ctr_n50_k3_p0.2_T3_ns3_s2] best solve_rate=0.0 mean_conflicts=102.423 @ep20 (208.2s)
+ wrote /home/yurenh2/rrog/runs/ckpt_color_full_none_ctr_n50_k3_p0.2_T3_ns3_s2.pt
+[full] LE n=300 fail_rate=1.00 | lambda1 SOLVED mean +nan (n=0) | UNSOLVED mean -0.9886 (n=300) | sep=+nan | AUROC(fail|lambda1)=nan | mean_lambda1=-0.9886
+[pe=none s2] deterministic solve_rate = 0.000 (n=150, K=16)
+ sigma pass@K lam-sel random perRoll AUROC(s|-lam)
+ 0.2 0.000 0.000 0.000 0.000 nan
+ wrote /home/yurenh2/rrog/runs/ptrm_color_full_none_ctr_n50_k3_p0.2_T3_ns3_s2.json
+===== seed=3 #3 full sub =====
+/home/yurenh2/rrog/diag/train_color.py:86: RuntimeWarning: invalid value encountered in divide
+ clus = np.where(deg > 1, tri / (deg * (deg - 1) / 2.0), 0.0)
+ep20 solve_rate=0.000 mean_conflicts=109.49
+ep40 solve_rate=0.253 mean_conflicts=27.50
+ep60 solve_rate=0.203 mean_conflicts=33.37
+ep80 solve_rate=0.227 mean_conflicts=59.83
+ep100 solve_rate=0.043 mean_conflicts=81.74
+ep120 solve_rate=0.080 mean_conflicts=75.93
+ep140 solve_rate=0.170 mean_conflicts=56.67
+ep150 solve_rate=0.087 mean_conflicts=85.03
+[color_full_sub_n50_k3_p0.2_T3_ns3_s3] best solve_rate=0.2533 mean_conflicts=27.497 @ep40 (152.2s)
+ wrote /home/yurenh2/rrog/runs/ckpt_color_full_sub_n50_k3_p0.2_T3_ns3_s3.pt
+/home/yurenh2/rrog/diag/train_color.py:86: RuntimeWarning: invalid value encountered in divide
+ clus = np.where(deg > 1, tri / (deg * (deg - 1) / 2.0), 0.0)
+[full] LE n=300 fail_rate=0.75 | lambda1 SOLVED mean -0.3797 (n=76) | UNSOLVED mean -0.1929 (n=224) | sep=+0.1869 | AUROC(fail|lambda1)=0.805 | mean_lambda1=-0.2402
+/home/yurenh2/rrog/diag/train_color.py:86: RuntimeWarning: invalid value encountered in divide
+ clus = np.where(deg > 1, tri / (deg * (deg - 1) / 2.0), 0.0)
+[pe=sub s3] deterministic solve_rate = 0.280 (n=150, K=16)
+ sigma pass@K lam-sel random perRoll AUROC(s|-lam)
+ 0.2 0.380 0.260 0.260 0.247 0.831
+ wrote /home/yurenh2/rrog/runs/ptrm_color_full_sub_n50_k3_p0.2_T3_ns3_s3.json
+===== seed=3 #4 full none --contract =====
+ep20 solve_rate=0.000 mean_conflicts=125.59
+ep40 solve_rate=0.000 mean_conflicts=118.58
+ep60 solve_rate=0.000 mean_conflicts=84.82
+ep80 solve_rate=0.000 mean_conflicts=65.44
+ep100 solve_rate=0.000 mean_conflicts=59.09
+ep120 solve_rate=0.000 mean_conflicts=61.66
+ep140 solve_rate=0.000 mean_conflicts=62.53
+ep150 solve_rate=0.000 mean_conflicts=63.31
+[color_full_none_ctr_n50_k3_p0.2_T3_ns3_s3] best solve_rate=0.0 mean_conflicts=125.593 @ep20 (206.2s)
+ wrote /home/yurenh2/rrog/runs/ckpt_color_full_none_ctr_n50_k3_p0.2_T3_ns3_s3.pt
+[full] LE n=300 fail_rate=1.00 | lambda1 SOLVED mean +nan (n=0) | UNSOLVED mean -0.5765 (n=300) | sep=+nan | AUROC(fail|lambda1)=nan | mean_lambda1=-0.5765
+[pe=none s3] deterministic solve_rate = 0.000 (n=150, K=16)
+ sigma pass@K lam-sel random perRoll AUROC(s|-lam)
+ 0.2 0.000 0.000 0.000 0.000 nan
+ wrote /home/yurenh2/rrog/runs/ptrm_color_full_none_ctr_n50_k3_p0.2_T3_ns3_s3.json
+===== seed=4 #3 full sub =====
+/home/yurenh2/rrog/diag/train_color.py:86: RuntimeWarning: invalid value encountered in divide
+ clus = np.where(deg > 1, tri / (deg * (deg - 1) / 2.0), 0.0)
+ep20 solve_rate=0.000 mean_conflicts=157.73
+ep40 solve_rate=0.140 mean_conflicts=24.79
+ep60 solve_rate=0.120 mean_conflicts=98.35
+ep80 solve_rate=0.623 mean_conflicts=4.32
+ep100 solve_rate=0.540 mean_conflicts=16.64
+ep120 solve_rate=0.380 mean_conflicts=41.04
+ep140 solve_rate=0.390 mean_conflicts=44.34
+ep150 solve_rate=0.387 mean_conflicts=43.48
+[color_full_sub_n50_k3_p0.2_T3_ns3_s4] best solve_rate=0.6233 mean_conflicts=4.323 @ep80 (152.1s)
+ wrote /home/yurenh2/rrog/runs/ckpt_color_full_sub_n50_k3_p0.2_T3_ns3_s4.pt
+/home/yurenh2/rrog/diag/train_color.py:86: RuntimeWarning: invalid value encountered in divide
+ clus = np.where(deg > 1, tri / (deg * (deg - 1) / 2.0), 0.0)
+[full] LE n=300 fail_rate=0.38 | lambda1 SOLVED mean -0.3913 (n=187) | UNSOLVED mean +0.0144 (n=113) | sep=+0.4057 | AUROC(fail|lambda1)=0.940 | mean_lambda1=-0.2385
+/home/yurenh2/rrog/diag/train_color.py:86: RuntimeWarning: invalid value encountered in divide
+ clus = np.where(deg > 1, tri / (deg * (deg - 1) / 2.0), 0.0)
+[pe=sub s4] deterministic solve_rate = 0.647 (n=150, K=16)
+ sigma pass@K lam-sel random perRoll AUROC(s|-lam)
+ 0.2 0.907 0.793 0.647 0.645 0.914
+ wrote /home/yurenh2/rrog/runs/ptrm_color_full_sub_n50_k3_p0.2_T3_ns3_s4.json
+===== seed=4 #4 full none --contract =====
+ep20 solve_rate=0.000 mean_conflicts=156.35
+ep40 solve_rate=0.000 mean_conflicts=121.15
+ep60 solve_rate=0.000 mean_conflicts=82.09
+ep80 solve_rate=0.000 mean_conflicts=80.90
+ep100 solve_rate=0.000 mean_conflicts=81.51
+ep120 solve_rate=0.000 mean_conflicts=80.79
+ep140 solve_rate=0.000 mean_conflicts=74.61
+ep150 solve_rate=0.000 mean_conflicts=73.39
+[color_full_none_ctr_n50_k3_p0.2_T3_ns3_s4] best solve_rate=0.0 mean_conflicts=156.353 @ep20 (208.3s)
+ wrote /home/yurenh2/rrog/runs/ckpt_color_full_none_ctr_n50_k3_p0.2_T3_ns3_s4.pt
+[full] LE n=300 fail_rate=1.00 | lambda1 SOLVED mean +nan (n=0) | UNSOLVED mean -0.7751 (n=300) | sep=+nan | AUROC(fail|lambda1)=nan | mean_lambda1=-0.7751
+[pe=none s4] deterministic solve_rate = 0.000 (n=150, K=16)
+ sigma pass@K lam-sel random perRoll AUROC(s|-lam)
+ 0.2 0.000 0.000 0.000 0.000 nan
+ wrote /home/yurenh2/rrog/runs/ptrm_color_full_none_ctr_n50_k3_p0.2_T3_ns3_s4.json
+===== AGGREGATE =====
+=== best solve_rate (deterministic, EMA) ===
+ ('1step', 'none', '-'): 0.293±0.062 (n=5)
+ ('full', 'gsn', '-'): 0.596±0.044 (n=5)
+ ('full', 'none', '-'): 0.457±0.087 (n=5)
+ ('full', 'none', 'ctr'): 0.000±0.000 (n=5)
+ ('full', 'rwse', '-'): 0.503±0.064 (n=5)
+ ('full', 'sub', '-'): 0.565±0.163 (n=5)
+=== LE AUROC(fail|lambda1) ===
+ ('1step', 'none', '-'): 0.821±0.036 (n=5)
+ ('full', 'gsn', '-'): 0.868±0.023 (n=5)
+ ('full', 'none', '-'): 0.868±0.032 (n=5)
+ ('full', 'none', 'ctr'): nan±nan (n=5)
+ ('full', 'rwse', '-'): 0.837±0.041 (n=5)
+ ('full', 'sub', '-'): 0.899±0.051 (n=5)
+=== LE mean_lambda1 (forced-contraction dose) ===
+ ('1step', 'none', '-'): -0.051±0.092 (n=5)
+ ('full', 'gsn', '-'): -0.190±0.051 (n=5)
+ ('full', 'none', '-'): -0.197±0.032 (n=5)
+ ('full', 'none', 'ctr'): -0.804±0.188 (n=5)
+ ('full', 'rwse', '-'): -0.011±0.075 (n=5)
+ ('full', 'sub', '-'): -0.220±0.028 (n=5)
+=== PTRM sigma=0.2 ===
+ ('full', 'gsn', '-'): det 0.595±0.039 (n=5) | pass@K 0.824±0.025 (n=5) | lambda-sel 0.697±0.040 (n=5) | AUROC 0.875±0.014 (n=5)
+ ('full', 'none', '-'): det 0.480±0.088 (n=5) | pass@K 0.739±0.113 (n=5) | lambda-sel 0.600±0.110 (n=5) | AUROC 0.853±0.023 (n=5)
+ ('full', 'none', 'ctr'): det 0.000±0.000 (n=5) | pass@K 0.000±0.000 (n=5) | lambda-sel 0.000±0.000 (n=5) | AUROC nan±nan (n=5)
+ ('full', 'rwse', '-'): det 0.521±0.065 (n=5) | pass@K 0.883±0.075 (n=5) | lambda-sel 0.736±0.088 (n=5) | AUROC 0.837±0.028 (n=5)
+ ('full', 'sub', '-'): det 0.581±0.165 (n=5) | pass@K 0.793±0.207 (n=5) | lambda-sel 0.685±0.218 (n=5) | AUROC 0.894±0.033 (n=5)
+done=2026-06-16T14:24:48-05:00
generated by cgit v1.2.3 (git 2.25.1) at 2026-06-30 22:31:33 +0000