1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
|
host=timan1.cs.illinois.edu gpu=0 start=2026-06-16T11:31:39-05:00
===== train full pe=none =====
ep20 solve_rate=0.000 mean_conflicts=122.82
ep40 solve_rate=0.167 mean_conflicts=60.87
ep60 solve_rate=0.187 mean_conflicts=64.07
ep80 solve_rate=0.313 mean_conflicts=23.74
ep100 solve_rate=0.093 mean_conflicts=77.11
ep120 solve_rate=0.060 mean_conflicts=69.50
ep140 solve_rate=0.113 mean_conflicts=38.00
ep150 solve_rate=0.103 mean_conflicts=39.66
[color_full_none_n50_k3_p0.2_T3_ns3_s0] best solve_rate=0.3133 mean_conflicts=23.737 @ep80 (152.8s)
wrote /home/yurenh2/rrog/runs/ckpt_color_full_none_n50_k3_p0.2_T3_ns3_s0.pt
===== train full pe=rwse =====
ep20 solve_rate=0.000 mean_conflicts=71.30
ep40 solve_rate=0.193 mean_conflicts=29.00
ep60 solve_rate=0.177 mean_conflicts=37.01
ep80 solve_rate=0.483 mean_conflicts=6.12
ep100 solve_rate=0.543 mean_conflicts=7.36
ep120 solve_rate=0.497 mean_conflicts=13.22
ep140 solve_rate=0.373 mean_conflicts=49.20
ep150 solve_rate=0.363 mean_conflicts=50.49
[color_full_rwse_n50_k3_p0.2_T3_ns3_s0] best solve_rate=0.5433 mean_conflicts=7.357 @ep100 (152.9s)
wrote /home/yurenh2/rrog/runs/ckpt_color_full_rwse_n50_k3_p0.2_T3_ns3_s0.pt
===== LE (full, both pe) =====
[full] LE n=300 fail_rate=0.69 | lambda1 SOLVED mean -0.2724 (n=94) | UNSOLVED mean -0.0750 (n=206) | sep=+0.1974 | AUROC(fail|lambda1)=0.823 | mean_lambda1=-0.1369
Traceback (most recent call last):
File "/home/yurenh2/rrog/diag/train_color.py", line 254, in <module>
main()
~~~~^^
File "/home/yurenh2/rrog/diag/train_color.py", line 202, in main
run_le(model, te, dev, c['n_sup'] * c['T'])
~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/yurenh2/rrog/diag/train_color.py", line 159, in run_le
col = model(xin, ei)[-1].argmax(-1)
~~~~~^^^^^^^^^
File "/home/yurenh2/miniconda3/lib/python3.13/site-packages/torch/nn/modules/module.py", line 1776, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
File "/home/yurenh2/miniconda3/lib/python3.13/site-packages/torch/nn/modules/module.py", line 1787, in _call_impl
return forward_call(*args, **kwargs)
File "/home/yurenh2/rrog/diag/train_color.py", line 106, in forward
h0 = self.lin_in(xin)
File "/home/yurenh2/miniconda3/lib/python3.13/site-packages/torch/nn/modules/module.py", line 1776, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
File "/home/yurenh2/miniconda3/lib/python3.13/site-packages/torch/nn/modules/module.py", line 1787, in _call_impl
return forward_call(*args, **kwargs)
File "/home/yurenh2/miniconda3/lib/python3.13/site-packages/torch/nn/modules/linear.py", line 134, in forward
return F.linear(input, self.weight, self.bias)
~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
RuntimeError: mat1 and mat2 shapes cannot be multiplied (50x8 and 24x128)
!! le rwse failed
===== PTRM noise + lambda-select (both pe) =====
--- pe=none ---
[pe=none] deterministic solve_rate = 0.327 (n=150, K=16)
sigma pass@K lam-sel random perRoll AUROC(s|-lam)
0.1 0.560 0.440 0.353 0.323 0.820
0.2 0.747 0.520 0.307 0.313 0.817
0.4 0.860 0.620 0.300 0.292 0.777
--- pe=rwse ---
[pe=rwse] deterministic solve_rate = 0.593 (n=150, K=16)
sigma pass@K lam-sel random perRoll AUROC(s|-lam)
0.1 0.840 0.720 0.593 0.595 0.850
0.2 0.927 0.767 0.540 0.579 0.838
0.4 0.973 0.860 0.540 0.509 0.831
done=2026-06-16T11:41:49-05:00
|
