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
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
|
[run] ogbg-molclintox view=gin compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molclintox --view gin --compute rrog-act --T 1 --n_sup 3 --hidden 128 --bs 128 --epochs 100 --eval_every 10 --agg_layers 5 --compute_layers 2 --seed 0 --lam_q 0.1 --halt_max_steps 8 --halt_min_steps 2 --halt_target loss --halt_loss_threshold 0.2 --halt_exploration_prob 0.1 --act_train_mode stream --q_warmup_epochs 0 --device cuda:1 --num_workers 0
ep10 val_rocauc=0.77209 val_adapt_rocauc=0.80519 adapt_steps=2.72 halt=0.34 train_steps=2.51
ep20 val_rocauc=0.78183 val_adapt_rocauc=0.80016 adapt_steps=2.06 halt=0.37 train_steps=2.64
ep30 val_rocauc=0.78255 val_adapt_rocauc=0.76576 adapt_steps=4.13 halt=0.34 train_steps=2.65
ep40 val_rocauc=0.94888 val_adapt_rocauc=0.93689 adapt_steps=2.55 halt=0.37 train_steps=2.28
ep50 val_rocauc=0.88011 val_adapt_rocauc=0.87381 adapt_steps=2.15 halt=0.38 train_steps=2.23
ep60 val_rocauc=0.95997 val_adapt_rocauc=0.96159 adapt_steps=2.24 halt=0.37 train_steps=2.55
ep70 val_rocauc=0.92487 val_adapt_rocauc=0.91931 adapt_steps=2.11 halt=0.39 train_steps=2.34
ep80 val_rocauc=0.93713 val_adapt_rocauc=0.92883 adapt_steps=2.45 halt=0.42 train_steps=2.15
ep90 val_rocauc=0.94286 val_adapt_rocauc=0.93582 adapt_steps=2.32 halt=0.41 train_steps=2.08
ep100 val_rocauc=0.93878 val_adapt_rocauc=0.92695 adapt_steps=2.21 halt=0.40 train_steps=2.15
[ogbg-molclintox_gin_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=60 val={'rocauc': 0.9599674972914409} test={'rocauc': 0.8033521009279533} adaptive={'rocauc': 0.8111476036562055} steps=2.7094594594594597
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molclintox_gin_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molclintox view=gine compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molclintox --view gine --compute rrog-act --T 1 --n_sup 3 --hidden 128 --bs 128 --epochs 100 --eval_every 10 --agg_layers 5 --compute_layers 2 --seed 0 --lam_q 0.1 --halt_max_steps 8 --halt_min_steps 2 --halt_target loss --halt_loss_threshold 0.2 --halt_exploration_prob 0.1 --act_train_mode stream --q_warmup_epochs 0 --device cuda:1 --num_workers 0
ep10 val_rocauc=0.76383 val_adapt_rocauc=0.77994 adapt_steps=2.01 halt=0.36 train_steps=2.43
ep20 val_rocauc=0.80870 val_adapt_rocauc=0.86788 adapt_steps=3.20 halt=0.36 train_steps=2.52
ep30 val_rocauc=0.75884 val_adapt_rocauc=0.80281 adapt_steps=2.47 halt=0.35 train_steps=2.59
ep40 val_rocauc=0.86243 val_adapt_rocauc=0.89579 adapt_steps=2.90 halt=0.34 train_steps=2.62
ep50 val_rocauc=0.89135 val_adapt_rocauc=0.86760 adapt_steps=2.41 halt=0.38 train_steps=2.29
ep60 val_rocauc=0.90485 val_adapt_rocauc=0.88818 adapt_steps=2.26 halt=0.40 train_steps=2.20
ep70 val_rocauc=0.93950 val_adapt_rocauc=0.93379 adapt_steps=2.16 halt=0.39 train_steps=2.27
ep80 val_rocauc=0.94562 val_adapt_rocauc=0.91967 adapt_steps=2.22 halt=0.40 train_steps=2.25
ep90 val_rocauc=0.93472 val_adapt_rocauc=0.91171 adapt_steps=2.22 halt=0.41 train_steps=2.02
ep100 val_rocauc=0.93612 val_adapt_rocauc=0.91308 adapt_steps=2.22 halt=0.41 train_steps=2.14
[ogbg-molclintox_gine_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=80 val={'rocauc': 0.9456195213237466} test={'rocauc': 0.9048100649914852} adaptive={'rocauc': 0.9068831891008932} steps=2.2972972972972974
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molclintox_gine_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molclintox view=gcn compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molclintox --view gcn --compute rrog-act --T 1 --n_sup 3 --hidden 128 --bs 128 --epochs 100 --eval_every 10 --agg_layers 5 --compute_layers 2 --seed 0 --lam_q 0.1 --halt_max_steps 8 --halt_min_steps 2 --halt_target loss --halt_loss_threshold 0.2 --halt_exploration_prob 0.1 --act_train_mode stream --q_warmup_epochs 0 --device cuda:1 --num_workers 0
ep10 val_rocauc=0.85640 val_adapt_rocauc=0.82455 adapt_steps=2.11 halt=0.38 train_steps=2.49
ep20 val_rocauc=0.82161 val_adapt_rocauc=0.86794 adapt_steps=2.25 halt=0.42 train_steps=1.95
ep30 val_rocauc=0.90791 val_adapt_rocauc=0.90606 adapt_steps=2.48 halt=0.40 train_steps=2.22
ep40 val_rocauc=0.87426 val_adapt_rocauc=0.86053 adapt_steps=2.28 halt=0.39 train_steps=2.23
ep50 val_rocauc=0.90999 val_adapt_rocauc=0.93044 adapt_steps=2.48 halt=0.40 train_steps=2.19
ep60 val_rocauc=0.92425 val_adapt_rocauc=0.92174 adapt_steps=2.33 halt=0.43 train_steps=1.93
ep70 val_rocauc=0.91453 val_adapt_rocauc=0.92876 adapt_steps=2.26 halt=0.42 train_steps=2.06
ep80 val_rocauc=0.94111 val_adapt_rocauc=0.92515 adapt_steps=2.18 halt=0.44 train_steps=1.91
ep90 val_rocauc=0.94603 val_adapt_rocauc=0.94386 adapt_steps=2.16 halt=0.43 train_steps=1.97
ep100 val_rocauc=0.93669 val_adapt_rocauc=0.93404 adapt_steps=2.24 halt=0.42 train_steps=1.93
[ogbg-molclintox_gcn_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=90 val={'rocauc': 0.9460307298335469} test={'rocauc': 0.8712221527126125} adaptive={'rocauc': 0.8801906301046121} steps=2.2027027027027026
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molclintox_gcn_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molclintox view=graphsage compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molclintox --view graphsage --compute rrog-act --T 1 --n_sup 3 --hidden 128 --bs 128 --epochs 100 --eval_every 10 --agg_layers 5 --compute_layers 2 --seed 0 --lam_q 0.1 --halt_max_steps 8 --halt_min_steps 2 --halt_target loss --halt_loss_threshold 0.2 --halt_exploration_prob 0.1 --act_train_mode stream --q_warmup_epochs 0 --device cuda:1 --num_workers 0
ep10 val_rocauc=0.79462 val_adapt_rocauc=0.82656 adapt_steps=2.22 halt=0.34 train_steps=2.52
ep20 val_rocauc=0.84462 val_adapt_rocauc=0.87055 adapt_steps=2.30 halt=0.38 train_steps=2.43
ep30 val_rocauc=0.75120 val_adapt_rocauc=0.82551 adapt_steps=2.07 halt=0.40 train_steps=2.12
ep40 val_rocauc=0.93718 val_adapt_rocauc=0.93679 adapt_steps=2.45 halt=0.39 train_steps=2.33
ep50 val_rocauc=0.94727 val_adapt_rocauc=0.94822 adapt_steps=2.40 halt=0.40 train_steps=2.11
ep60 val_rocauc=0.95164 val_adapt_rocauc=0.94490 adapt_steps=2.51 halt=0.40 train_steps=2.18
ep70 val_rocauc=0.87261 val_adapt_rocauc=0.87261 adapt_steps=2.23 halt=0.40 train_steps=2.12
ep80 val_rocauc=0.89686 val_adapt_rocauc=0.89115 adapt_steps=2.20 halt=0.42 train_steps=2.05
ep90 val_rocauc=0.88179 val_adapt_rocauc=0.87664 adapt_steps=2.16 halt=0.41 train_steps=2.02
ep100 val_rocauc=0.89555 val_adapt_rocauc=0.89160 adapt_steps=2.14 halt=0.41 train_steps=2.08
[ogbg-molclintox_graphsage_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=60 val={'rocauc': 0.9516390886109196} test={'rocauc': 0.88216296527995} adaptive={'rocauc': 0.8780280123727106} steps=2.4324324324324325
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molclintox_graphsage_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molclintox view=gatv2 compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molclintox --view gatv2 --compute rrog-act --T 1 --n_sup 3 --hidden 128 --bs 128 --epochs 100 --eval_every 10 --agg_layers 5 --compute_layers 2 --seed 0 --lam_q 0.1 --halt_max_steps 8 --halt_min_steps 2 --halt_target loss --halt_loss_threshold 0.2 --halt_exploration_prob 0.1 --act_train_mode stream --q_warmup_epochs 0 --device cuda:1 --num_workers 0
ep10 val_rocauc=0.70967 val_adapt_rocauc=0.85199 adapt_steps=2.37 halt=0.37 train_steps=2.39
ep20 val_rocauc=0.89092 val_adapt_rocauc=0.92549 adapt_steps=2.22 halt=0.44 train_steps=1.89
ep30 val_rocauc=0.81332 val_adapt_rocauc=0.83200 adapt_steps=2.01 halt=0.45 train_steps=1.79
ep40 val_rocauc=0.80582 val_adapt_rocauc=0.83143 adapt_steps=2.14 halt=0.40 train_steps=2.12
ep50 val_rocauc=0.92910 val_adapt_rocauc=0.93248 adapt_steps=2.32 halt=0.41 train_steps=2.22
ep60 val_rocauc=0.92482 val_adapt_rocauc=0.93468 adapt_steps=2.32 halt=0.41 train_steps=2.17
ep70 val_rocauc=0.92105 val_adapt_rocauc=0.93623 adapt_steps=2.38 halt=0.41 train_steps=2.15
ep80 val_rocauc=0.92617 val_adapt_rocauc=0.93959 adapt_steps=2.31 halt=0.41 train_steps=2.12
ep90 val_rocauc=0.94562 val_adapt_rocauc=0.95085 adapt_steps=2.28 halt=0.41 train_steps=2.20
ep100 val_rocauc=0.95068 val_adapt_rocauc=0.95672 adapt_steps=2.28 halt=0.40 train_steps=2.14
[ogbg-molclintox_gatv2_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=100 val={'rocauc': 0.9506763189861781} test={'rocauc': 0.8617905675459633} adaptive={'rocauc': 0.834407256803253} steps=2.3378378378378377
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molclintox_gatv2_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molclintox view=graphconv compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molclintox --view graphconv --compute rrog-act --T 1 --n_sup 3 --hidden 128 --bs 128 --epochs 100 --eval_every 10 --agg_layers 5 --compute_layers 2 --seed 0 --lam_q 0.1 --halt_max_steps 8 --halt_min_steps 2 --halt_target loss --halt_loss_threshold 0.2 --halt_exploration_prob 0.1 --act_train_mode stream --q_warmup_epochs 0 --device cuda:1 --num_workers 0
ep10 val_rocauc=0.69917 val_adapt_rocauc=0.72403 adapt_steps=2.24 halt=0.39 train_steps=2.26
ep20 val_rocauc=0.81950 val_adapt_rocauc=0.82965 adapt_steps=2.00 halt=0.34 train_steps=2.75
ep30 val_rocauc=0.80323 val_adapt_rocauc=0.78038 adapt_steps=2.73 halt=0.34 train_steps=2.70
ep40 val_rocauc=0.89633 val_adapt_rocauc=0.85819 adapt_steps=2.17 halt=0.37 train_steps=2.38
ep50 val_rocauc=0.93302 val_adapt_rocauc=0.93985 adapt_steps=2.71 halt=0.38 train_steps=2.37
ep60 val_rocauc=0.94116 val_adapt_rocauc=0.94729 adapt_steps=2.39 halt=0.41 train_steps=2.13
ep70 val_rocauc=0.95746 val_adapt_rocauc=0.93137 adapt_steps=2.28 halt=0.41 train_steps=2.12
ep80 val_rocauc=0.94965 val_adapt_rocauc=0.93652 adapt_steps=2.24 halt=0.41 train_steps=2.13
ep90 val_rocauc=0.95186 val_adapt_rocauc=0.93864 adapt_steps=2.24 halt=0.41 train_steps=2.05
ep100 val_rocauc=0.95550 val_adapt_rocauc=0.93900 adapt_steps=2.28 halt=0.41 train_steps=2.03
[ogbg-molclintox_graphconv_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=70 val={'rocauc': 0.9574583866837387} test={'rocauc': 0.881274111145866} adaptive={'rocauc': 0.8484968546901609} steps=2.4121621621621623
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molclintox_graphconv_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molclintox view=transformer compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molclintox --view transformer --compute rrog-act --T 1 --n_sup 3 --hidden 128 --bs 128 --epochs 100 --eval_every 10 --agg_layers 5 --compute_layers 2 --seed 0 --lam_q 0.1 --halt_max_steps 8 --halt_min_steps 2 --halt_target loss --halt_loss_threshold 0.2 --halt_exploration_prob 0.1 --act_train_mode stream --q_warmup_epochs 0 --device cuda:1 --num_workers 0
ep10 val_rocauc=0.66344 val_adapt_rocauc=0.75035 adapt_steps=2.91 halt=0.33 train_steps=2.69
ep20 val_rocauc=0.67594 val_adapt_rocauc=0.72203 adapt_steps=2.06 halt=0.39 train_steps=2.32
ep30 val_rocauc=0.85214 val_adapt_rocauc=0.88626 adapt_steps=2.31 halt=0.40 train_steps=2.06
ep40 val_rocauc=0.76707 val_adapt_rocauc=0.81695 adapt_steps=2.36 halt=0.38 train_steps=2.22
ep50 val_rocauc=0.76833 val_adapt_rocauc=0.77783 adapt_steps=2.26 halt=0.40 train_steps=2.13
ep60 val_rocauc=0.79067 val_adapt_rocauc=0.78862 adapt_steps=2.30 halt=0.43 train_steps=2.04
ep70 val_rocauc=0.79816 val_adapt_rocauc=0.79690 adapt_steps=2.20 halt=0.41 train_steps=1.99
ep80 val_rocauc=0.78835 val_adapt_rocauc=0.78999 adapt_steps=2.22 halt=0.41 train_steps=2.03
ep90 val_rocauc=0.78740 val_adapt_rocauc=0.78532 adapt_steps=2.21 halt=0.43 train_steps=1.98
ep100 val_rocauc=0.78905 val_adapt_rocauc=0.78792 adapt_steps=2.24 halt=0.42 train_steps=1.98
[ogbg-molclintox_transformer_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=30 val={'rocauc': 0.8521414031977412} test={'rocauc': 0.7332707746846001} adaptive={'rocauc': 0.7908464532721649} steps=3.2567567567567566
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molclintox_transformer_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molclintox view=pna compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molclintox --view pna --compute rrog-act --T 1 --n_sup 3 --hidden 128 --bs 128 --epochs 100 --eval_every 10 --agg_layers 5 --compute_layers 2 --seed 0 --lam_q 0.1 --halt_max_steps 8 --halt_min_steps 2 --halt_target loss --halt_loss_threshold 0.2 --halt_exploration_prob 0.1 --act_train_mode stream --q_warmup_epochs 0 --device cuda:1 --num_workers 0
/orion/u/oscarwan/rrog-gnn-runner/.venv/lib/python3.13/site-packages/torch_geometric/utils/_scatter.py:91: UserWarning: The usage of `scatter(reduce='min')` can be accelerated via the 'torch-scatter' package, but it was not found
warnings.warn(
/orion/u/oscarwan/rrog-gnn-runner/.venv/lib/python3.13/site-packages/torch_geometric/utils/_scatter.py:91: UserWarning: The usage of `scatter(reduce='max')` can be accelerated via the 'torch-scatter' package, but it was not found
warnings.warn(
ep10 val_rocauc=0.68589 val_adapt_rocauc=0.92664 adapt_steps=2.16 halt=0.41 train_steps=1.91
/orion/u/oscarwan/rrog-gnn-runner/.venv/lib/python3.13/site-packages/torch_geometric/utils/_scatter.py:91: UserWarning: The usage of `scatter(reduce='min')` can be accelerated via the 'torch-scatter' package, but it was not found
warnings.warn(
/orion/u/oscarwan/rrog-gnn-runner/.venv/lib/python3.13/site-packages/torch_geometric/utils/_scatter.py:91: UserWarning: The usage of `scatter(reduce='max')` can be accelerated via the 'torch-scatter' package, but it was not found
warnings.warn(
ep20 val_rocauc=0.85293 val_adapt_rocauc=0.85189 adapt_steps=2.44 halt=0.35 train_steps=2.56
/orion/u/oscarwan/rrog-gnn-runner/.venv/lib/python3.13/site-packages/torch_geometric/utils/_scatter.py:91: UserWarning: The usage of `scatter(reduce='min')` can be accelerated via the 'torch-scatter' package, but it was not found
warnings.warn(
/orion/u/oscarwan/rrog-gnn-runner/.venv/lib/python3.13/site-packages/torch_geometric/utils/_scatter.py:91: UserWarning: The usage of `scatter(reduce='max')` can be accelerated via the 'torch-scatter' package, but it was not found
warnings.warn(
ep30 val_rocauc=0.77476 val_adapt_rocauc=0.82179 adapt_steps=2.44 halt=0.28 train_steps=3.08
/orion/u/oscarwan/rrog-gnn-runner/.venv/lib/python3.13/site-packages/torch_geometric/utils/_scatter.py:91: UserWarning: The usage of `scatter(reduce='min')` can be accelerated via the 'torch-scatter' package, but it was not found
warnings.warn(
/orion/u/oscarwan/rrog-gnn-runner/.venv/lib/python3.13/site-packages/torch_geometric/utils/_scatter.py:91: UserWarning: The usage of `scatter(reduce='max')` can be accelerated via the 'torch-scatter' package, but it was not found
warnings.warn(
ep40 val_rocauc=0.87414 val_adapt_rocauc=0.89831 adapt_steps=2.53 halt=0.30 train_steps=2.83
/orion/u/oscarwan/rrog-gnn-runner/.venv/lib/python3.13/site-packages/torch_geometric/utils/_scatter.py:91: UserWarning: The usage of `scatter(reduce='min')` can be accelerated via the 'torch-scatter' package, but it was not found
warnings.warn(
/orion/u/oscarwan/rrog-gnn-runner/.venv/lib/python3.13/site-packages/torch_geometric/utils/_scatter.py:91: UserWarning: The usage of `scatter(reduce='max')` can be accelerated via the 'torch-scatter' package, but it was not found
warnings.warn(
ep50 val_rocauc=0.86257 val_adapt_rocauc=0.88010 adapt_steps=2.60 halt=0.32 train_steps=2.83
/orion/u/oscarwan/rrog-gnn-runner/.venv/lib/python3.13/site-packages/torch_geometric/utils/_scatter.py:91: UserWarning: The usage of `scatter(reduce='min')` can be accelerated via the 'torch-scatter' package, but it was not found
warnings.warn(
/orion/u/oscarwan/rrog-gnn-runner/.venv/lib/python3.13/site-packages/torch_geometric/utils/_scatter.py:91: UserWarning: The usage of `scatter(reduce='max')` can be accelerated via the 'torch-scatter' package, but it was not found
warnings.warn(
ep60 val_rocauc=0.71890 val_adapt_rocauc=0.77423 adapt_steps=3.07 halt=0.36 train_steps=2.45
/orion/u/oscarwan/rrog-gnn-runner/.venv/lib/python3.13/site-packages/torch_geometric/utils/_scatter.py:91: UserWarning: The usage of `scatter(reduce='min')` can be accelerated via the 'torch-scatter' package, but it was not found
warnings.warn(
/orion/u/oscarwan/rrog-gnn-runner/.venv/lib/python3.13/site-packages/torch_geometric/utils/_scatter.py:91: UserWarning: The usage of `scatter(reduce='max')` can be accelerated via the 'torch-scatter' package, but it was not found
warnings.warn(
ep70 val_rocauc=0.80193 val_adapt_rocauc=0.81742 adapt_steps=2.28 halt=0.37 train_steps=2.40
/orion/u/oscarwan/rrog-gnn-runner/.venv/lib/python3.13/site-packages/torch_geometric/utils/_scatter.py:91: UserWarning: The usage of `scatter(reduce='min')` can be accelerated via the 'torch-scatter' package, but it was not found
warnings.warn(
/orion/u/oscarwan/rrog-gnn-runner/.venv/lib/python3.13/site-packages/torch_geometric/utils/_scatter.py:91: UserWarning: The usage of `scatter(reduce='max')` can be accelerated via the 'torch-scatter' package, but it was not found
warnings.warn(
ep80 val_rocauc=0.82369 val_adapt_rocauc=0.84324 adapt_steps=2.29 halt=0.38 train_steps=2.34
/orion/u/oscarwan/rrog-gnn-runner/.venv/lib/python3.13/site-packages/torch_geometric/utils/_scatter.py:91: UserWarning: The usage of `scatter(reduce='min')` can be accelerated via the 'torch-scatter' package, but it was not found
warnings.warn(
/orion/u/oscarwan/rrog-gnn-runner/.venv/lib/python3.13/site-packages/torch_geometric/utils/_scatter.py:91: UserWarning: The usage of `scatter(reduce='max')` can be accelerated via the 'torch-scatter' package, but it was not found
warnings.warn(
ep90 val_rocauc=0.86654 val_adapt_rocauc=0.87053 adapt_steps=2.28 halt=0.41 train_steps=2.17
/orion/u/oscarwan/rrog-gnn-runner/.venv/lib/python3.13/site-packages/torch_geometric/utils/_scatter.py:91: UserWarning: The usage of `scatter(reduce='min')` can be accelerated via the 'torch-scatter' package, but it was not found
warnings.warn(
/orion/u/oscarwan/rrog-gnn-runner/.venv/lib/python3.13/site-packages/torch_geometric/utils/_scatter.py:91: UserWarning: The usage of `scatter(reduce='max')` can be accelerated via the 'torch-scatter' package, but it was not found
warnings.warn(
ep100 val_rocauc=0.86785 val_adapt_rocauc=0.86774 adapt_steps=2.23 halt=0.40 train_steps=2.14
[ogbg-molclintox_pna_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=40 val={'rocauc': 0.874143110410716} test={'rocauc': 0.8539203419872797} adaptive={'rocauc': 0.8613908872901679} steps=3.675675675675676
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molclintox_pna_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molclintox view=gen compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molclintox --view gen --compute rrog-act --T 1 --n_sup 3 --hidden 128 --bs 128 --epochs 100 --eval_every 10 --agg_layers 5 --compute_layers 2 --seed 0 --lam_q 0.1 --halt_max_steps 8 --halt_min_steps 2 --halt_target loss --halt_loss_threshold 0.2 --halt_exploration_prob 0.1 --act_train_mode stream --q_warmup_epochs 0 --device cuda:1 --num_workers 0
ep10 val_rocauc=0.84544 val_adapt_rocauc=0.85771 adapt_steps=2.57 halt=0.29 train_steps=2.87
ep20 val_rocauc=0.95110 val_adapt_rocauc=0.91548 adapt_steps=3.41 halt=0.33 train_steps=2.66
ep30 val_rocauc=0.83528 val_adapt_rocauc=0.89928 adapt_steps=2.29 halt=0.31 train_steps=2.91
ep40 val_rocauc=0.76843 val_adapt_rocauc=0.80352 adapt_steps=2.51 halt=0.34 train_steps=2.56
ep50 val_rocauc=0.80865 val_adapt_rocauc=0.85913 adapt_steps=2.77 halt=0.36 train_steps=2.47
ep60 val_rocauc=0.88044 val_adapt_rocauc=0.89659 adapt_steps=2.14 halt=0.42 train_steps=2.05
ep70 val_rocauc=0.92773 val_adapt_rocauc=0.93695 adapt_steps=2.28 halt=0.40 train_steps=2.12
ep80 val_rocauc=0.96265 val_adapt_rocauc=0.95789 adapt_steps=2.37 halt=0.39 train_steps=2.27
ep90 val_rocauc=0.95997 val_adapt_rocauc=0.95404 adapt_steps=2.19 halt=0.40 train_steps=2.22
ep100 val_rocauc=0.95236 val_adapt_rocauc=0.95381 adapt_steps=2.16 halt=0.40 train_steps=2.06
[ogbg-molclintox_gen_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=80 val={'rocauc': 0.962652253849437} test={'rocauc': 0.7993926597852152} adaptive={'rocauc': 0.8024780175859313} steps=2.52027027027027
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molclintox_gen_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molclintox view=film compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molclintox --view film --compute rrog-act --T 1 --n_sup 3 --hidden 128 --bs 128 --epochs 100 --eval_every 10 --agg_layers 5 --compute_layers 2 --seed 0 --lam_q 0.1 --halt_max_steps 8 --halt_min_steps 2 --halt_target loss --halt_loss_threshold 0.2 --halt_exploration_prob 0.1 --act_train_mode stream --q_warmup_epochs 0 --device cuda:1 --num_workers 0
ep10 val_rocauc=0.85317 val_adapt_rocauc=0.88605 adapt_steps=2.39 halt=0.28 train_steps=3.00
ep20 val_rocauc=0.81828 val_adapt_rocauc=0.82320 adapt_steps=2.45 halt=0.27 train_steps=3.23
ep30 val_rocauc=0.84258 val_adapt_rocauc=0.95090 adapt_steps=2.46 halt=0.34 train_steps=2.48
ep40 val_rocauc=0.75162 val_adapt_rocauc=0.79277 adapt_steps=2.35 halt=0.29 train_steps=3.14
ep50 val_rocauc=0.78711 val_adapt_rocauc=0.87848 adapt_steps=2.74 halt=0.28 train_steps=3.02
ep60 val_rocauc=0.81069 val_adapt_rocauc=0.91211 adapt_steps=2.38 halt=0.33 train_steps=2.62
ep70 val_rocauc=0.81552 val_adapt_rocauc=0.90454 adapt_steps=2.36 halt=0.33 train_steps=2.65
ep80 val_rocauc=0.82161 val_adapt_rocauc=0.91004 adapt_steps=2.32 halt=0.35 train_steps=2.57
ep90 val_rocauc=0.81750 val_adapt_rocauc=0.88712 adapt_steps=2.42 halt=0.34 train_steps=2.62
ep100 val_rocauc=0.82991 val_adapt_rocauc=0.90417 adapt_steps=2.39 halt=0.32 train_steps=2.82
[ogbg-molclintox_film_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=10 val={'rocauc': 0.8531673725335697} test={'rocauc': 0.8229512042539882} adaptive={'rocauc': 0.8409142251416258} steps=3.5675675675675675
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molclintox_film_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molclintox view=resgated compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molclintox --view resgated --compute rrog-act --T 1 --n_sup 3 --hidden 128 --bs 128 --epochs 100 --eval_every 10 --agg_layers 5 --compute_layers 2 --seed 0 --lam_q 0.1 --halt_max_steps 8 --halt_min_steps 2 --halt_target loss --halt_loss_threshold 0.2 --halt_exploration_prob 0.1 --act_train_mode stream --q_warmup_epochs 0 --device cuda:1 --num_workers 0
ep10 val_rocauc=0.74252 val_adapt_rocauc=0.79094 adapt_steps=2.14 halt=0.36 train_steps=2.48
ep20 val_rocauc=0.81568 val_adapt_rocauc=0.85910 adapt_steps=2.39 halt=0.35 train_steps=2.46
ep30 val_rocauc=0.84840 val_adapt_rocauc=0.92310 adapt_steps=2.11 halt=0.41 train_steps=2.09
ep40 val_rocauc=0.95769 val_adapt_rocauc=0.94539 adapt_steps=2.16 halt=0.43 train_steps=1.87
ep50 val_rocauc=0.92392 val_adapt_rocauc=0.94729 adapt_steps=2.65 halt=0.37 train_steps=2.31
ep60 val_rocauc=0.92694 val_adapt_rocauc=0.91922 adapt_steps=2.72 halt=0.37 train_steps=2.49
ep70 val_rocauc=0.90361 val_adapt_rocauc=0.89307 adapt_steps=2.32 halt=0.43 train_steps=1.98
ep80 val_rocauc=0.91214 val_adapt_rocauc=0.90406 adapt_steps=2.36 halt=0.41 train_steps=2.09
ep90 val_rocauc=0.90185 val_adapt_rocauc=0.89822 adapt_steps=2.32 halt=0.42 train_steps=2.04
ep100 val_rocauc=0.89869 val_adapt_rocauc=0.89844 adapt_steps=2.29 halt=0.42 train_steps=1.99
[ogbg-molclintox_resgated_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=40 val={'rocauc': 0.957693949243245} test={'rocauc': 0.8777291210509853} adaptive={'rocauc': 0.8950795885031105} steps=2.2567567567567566
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molclintox_resgated_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molclintox view=tag compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molclintox --view tag --compute rrog-act --T 1 --n_sup 3 --hidden 128 --bs 128 --epochs 100 --eval_every 10 --agg_layers 5 --compute_layers 2 --seed 0 --lam_q 0.1 --halt_max_steps 8 --halt_min_steps 2 --halt_target loss --halt_loss_threshold 0.2 --halt_exploration_prob 0.1 --act_train_mode stream --q_warmup_epochs 0 --device cuda:1 --num_workers 0
ep10 val_rocauc=0.82868 val_adapt_rocauc=0.84050 adapt_steps=2.43 halt=0.39 train_steps=2.17
ep20 val_rocauc=0.68167 val_adapt_rocauc=0.82555 adapt_steps=2.24 halt=0.39 train_steps=2.29
ep30 val_rocauc=0.90011 val_adapt_rocauc=0.87969 adapt_steps=2.48 halt=0.35 train_steps=2.43
ep40 val_rocauc=0.87408 val_adapt_rocauc=0.90534 adapt_steps=2.23 halt=0.38 train_steps=2.23
ep50 val_rocauc=0.86302 val_adapt_rocauc=0.87087 adapt_steps=2.50 halt=0.42 train_steps=2.01
ep60 val_rocauc=0.88360 val_adapt_rocauc=0.90039 adapt_steps=2.14 halt=0.41 train_steps=2.16
ep70 val_rocauc=0.86734 val_adapt_rocauc=0.88072 adapt_steps=2.36 halt=0.42 train_steps=1.95
ep80 val_rocauc=0.87492 val_adapt_rocauc=0.87176 adapt_steps=2.22 halt=0.42 train_steps=1.94
ep90 val_rocauc=0.87822 val_adapt_rocauc=0.87936 adapt_steps=2.20 halt=0.42 train_steps=2.04
ep100 val_rocauc=0.87463 val_adapt_rocauc=0.87752 adapt_steps=2.26 halt=0.42 train_steps=1.91
[ogbg-molclintox_tag_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=30 val={'rocauc': 0.9001050592599888} test={'rocauc': 0.8008871163938415} adaptive={'rocauc': 0.8083576616967296} steps=2.5945945945945947
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molclintox_tag_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molclintox view=sgc compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molclintox --view sgc --compute rrog-act --T 1 --n_sup 3 --hidden 128 --bs 128 --epochs 100 --eval_every 10 --agg_layers 5 --compute_layers 2 --seed 0 --lam_q 0.1 --halt_max_steps 8 --halt_min_steps 2 --halt_target loss --halt_loss_threshold 0.2 --halt_exploration_prob 0.1 --act_train_mode stream --q_warmup_epochs 0 --device cuda:1 --num_workers 0
ep10 val_rocauc=0.75045 val_adapt_rocauc=0.76256 adapt_steps=2.63 halt=0.35 train_steps=2.42
ep20 val_rocauc=0.84107 val_adapt_rocauc=0.82692 adapt_steps=2.05 halt=0.39 train_steps=2.28
ep30 val_rocauc=0.82797 val_adapt_rocauc=0.80820 adapt_steps=2.16 halt=0.40 train_steps=2.17
ep40 val_rocauc=0.91055 val_adapt_rocauc=0.89387 adapt_steps=2.19 halt=0.43 train_steps=1.88
ep50 val_rocauc=0.89428 val_adapt_rocauc=0.87914 adapt_steps=2.42 halt=0.38 train_steps=2.22
ep60 val_rocauc=0.89904 val_adapt_rocauc=0.88187 adapt_steps=2.34 halt=0.43 train_steps=1.92
ep70 val_rocauc=0.91524 val_adapt_rocauc=0.90461 adapt_steps=2.33 halt=0.43 train_steps=1.88
ep80 val_rocauc=0.91526 val_adapt_rocauc=0.89527 adapt_steps=2.31 halt=0.43 train_steps=1.99
ep90 val_rocauc=0.91339 val_adapt_rocauc=0.90184 adapt_steps=2.32 halt=0.44 train_steps=1.94
ep100 val_rocauc=0.91057 val_adapt_rocauc=0.90055 adapt_steps=2.30 halt=0.42 train_steps=2.02
[ogbg-molclintox_sgc_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=80 val={'rocauc': 0.9152647821661906} test={'rocauc': 0.8882589580509506} adaptive={'rocauc': 0.8904589371980676} steps=2.25
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molclintox_sgc_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molclintox view=cheb compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molclintox --view cheb --compute rrog-act --T 1 --n_sup 3 --hidden 128 --bs 128 --epochs 100 --eval_every 10 --agg_layers 5 --compute_layers 2 --seed 0 --lam_q 0.1 --halt_max_steps 8 --halt_min_steps 2 --halt_target loss --halt_loss_threshold 0.2 --halt_exploration_prob 0.1 --act_train_mode stream --q_warmup_epochs 0 --device cuda:1 --num_workers 0
ep10 val_rocauc=0.78798 val_adapt_rocauc=0.83382 adapt_steps=2.20 halt=0.39 train_steps=2.27
ep20 val_rocauc=0.71633 val_adapt_rocauc=0.74377 adapt_steps=2.78 halt=0.42 train_steps=1.99
ep30 val_rocauc=0.89270 val_adapt_rocauc=0.90791 adapt_steps=2.54 halt=0.41 train_steps=1.95
ep40 val_rocauc=0.91721 val_adapt_rocauc=0.93704 adapt_steps=2.25 halt=0.42 train_steps=1.89
ep50 val_rocauc=0.88995 val_adapt_rocauc=0.88075 adapt_steps=2.22 halt=0.40 train_steps=2.14
ep60 val_rocauc=0.87245 val_adapt_rocauc=0.86085 adapt_steps=2.17 halt=0.44 train_steps=1.96
ep70 val_rocauc=0.89380 val_adapt_rocauc=0.88259 adapt_steps=2.09 halt=0.42 train_steps=1.94
ep80 val_rocauc=0.89714 val_adapt_rocauc=0.89109 adapt_steps=2.17 halt=0.44 train_steps=1.92
ep90 val_rocauc=0.90792 val_adapt_rocauc=0.89461 adapt_steps=2.18 halt=0.43 train_steps=1.91
ep100 val_rocauc=0.91072 val_adapt_rocauc=0.90068 adapt_steps=2.14 halt=0.41 train_steps=1.96
[ogbg-molclintox_cheb_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=40 val={'rocauc': 0.9172116615778587} test={'rocauc': 0.906584297779168} adaptive={'rocauc': 0.9058335939943698} steps=2.97972972972973
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molclintox_cheb_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molclintox view=arma compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molclintox --view arma --compute rrog-act --T 1 --n_sup 3 --hidden 128 --bs 128 --epochs 100 --eval_every 10 --agg_layers 5 --compute_layers 2 --seed 0 --lam_q 0.1 --halt_max_steps 8 --halt_min_steps 2 --halt_target loss --halt_loss_threshold 0.2 --halt_exploration_prob 0.1 --act_train_mode stream --q_warmup_epochs 0 --device cuda:1 --num_workers 0
ep10 val_rocauc=0.81318 val_adapt_rocauc=0.80930 adapt_steps=2.37 halt=0.35 train_steps=2.35
ep20 val_rocauc=0.69804 val_adapt_rocauc=0.82638 adapt_steps=2.15 halt=0.31 train_steps=2.87
ep30 val_rocauc=0.94232 val_adapt_rocauc=0.92281 adapt_steps=2.83 halt=0.39 train_steps=2.09
ep40 val_rocauc=0.86469 val_adapt_rocauc=0.88327 adapt_steps=2.67 halt=0.36 train_steps=2.54
ep50 val_rocauc=0.89535 val_adapt_rocauc=0.90887 adapt_steps=2.19 halt=0.37 train_steps=2.46
ep60 val_rocauc=0.92823 val_adapt_rocauc=0.93065 adapt_steps=2.24 halt=0.41 train_steps=2.09
ep70 val_rocauc=0.93024 val_adapt_rocauc=0.93498 adapt_steps=2.20 halt=0.43 train_steps=1.91
ep80 val_rocauc=0.92742 val_adapt_rocauc=0.93392 adapt_steps=2.17 halt=0.42 train_steps=2.07
ep90 val_rocauc=0.93094 val_adapt_rocauc=0.94491 adapt_steps=2.17 halt=0.45 train_steps=1.81
ep100 val_rocauc=0.93728 val_adapt_rocauc=0.94060 adapt_steps=2.16 halt=0.41 train_steps=2.07
[ogbg-molclintox_arma_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=30 val={'rocauc': 0.942323287041597} test={'rocauc': 0.8458372432488792} adaptive={'rocauc': 0.8651226844611268} steps=2.722972972972973
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molclintox_arma_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molclintox view=mf compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molclintox --view mf --compute rrog-act --T 1 --n_sup 3 --hidden 128 --bs 128 --epochs 100 --eval_every 10 --agg_layers 5 --compute_layers 2 --seed 0 --lam_q 0.1 --halt_max_steps 8 --halt_min_steps 2 --halt_target loss --halt_loss_threshold 0.2 --halt_exploration_prob 0.1 --act_train_mode stream --q_warmup_epochs 0 --device cuda:1 --num_workers 0
ep10 val_rocauc=0.93105 val_adapt_rocauc=0.89417 adapt_steps=2.82 halt=0.32 train_steps=2.64
ep20 val_rocauc=0.81868 val_adapt_rocauc=0.81868 adapt_steps=2.54 halt=0.35 train_steps=2.46
ep30 val_rocauc=0.80286 val_adapt_rocauc=0.78639 adapt_steps=2.11 halt=0.37 train_steps=2.59
ep40 val_rocauc=0.80906 val_adapt_rocauc=0.83030 adapt_steps=2.18 halt=0.36 train_steps=2.49
ep50 val_rocauc=0.85998 val_adapt_rocauc=0.83241 adapt_steps=2.39 halt=0.40 train_steps=2.20
ep60 val_rocauc=0.90338 val_adapt_rocauc=0.86791 adapt_steps=2.18 halt=0.42 train_steps=2.08
ep70 val_rocauc=0.94162 val_adapt_rocauc=0.91534 adapt_steps=2.16 halt=0.42 train_steps=2.04
ep80 val_rocauc=0.89972 val_adapt_rocauc=0.89221 adapt_steps=2.14 halt=0.43 train_steps=2.03
ep90 val_rocauc=0.91851 val_adapt_rocauc=0.90667 adapt_steps=2.11 halt=0.42 train_steps=2.01
ep100 val_rocauc=0.92030 val_adapt_rocauc=0.90645 adapt_steps=2.14 halt=0.41 train_steps=2.03
[ogbg-molclintox_mf_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=70 val={'rocauc': 0.9416239863422963} test={'rocauc': 0.852654398220554} adaptive={'rocauc': 0.8698067632850242} steps=2.5675675675675675
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molclintox_mf_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molclintox view=appnp compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molclintox --view appnp --compute rrog-act --T 1 --n_sup 3 --hidden 128 --bs 128 --epochs 100 --eval_every 10 --agg_layers 5 --compute_layers 2 --seed 0 --lam_q 0.1 --halt_max_steps 8 --halt_min_steps 2 --halt_target loss --halt_loss_threshold 0.2 --halt_exploration_prob 0.1 --act_train_mode stream --q_warmup_epochs 0 --device cuda:1 --num_workers 0
ep10 val_rocauc=0.71982 val_adapt_rocauc=0.77487 adapt_steps=2.08 halt=0.44 train_steps=1.89
ep20 val_rocauc=0.74144 val_adapt_rocauc=0.75935 adapt_steps=2.03 halt=0.46 train_steps=1.69
ep30 val_rocauc=0.87530 val_adapt_rocauc=0.86698 adapt_steps=2.09 halt=0.45 train_steps=1.73
ep40 val_rocauc=0.85439 val_adapt_rocauc=0.85187 adapt_steps=2.16 halt=0.42 train_steps=1.92
ep50 val_rocauc=0.85586 val_adapt_rocauc=0.88542 adapt_steps=2.04 halt=0.44 train_steps=1.83
ep60 val_rocauc=0.91564 val_adapt_rocauc=0.92289 adapt_steps=2.16 halt=0.45 train_steps=1.84
ep70 val_rocauc=0.85864 val_adapt_rocauc=0.90628 adapt_steps=2.09 halt=0.45 train_steps=1.71
ep80 val_rocauc=0.85984 val_adapt_rocauc=0.91203 adapt_steps=2.01 halt=0.44 train_steps=1.90
ep90 val_rocauc=0.86275 val_adapt_rocauc=0.89833 adapt_steps=2.03 halt=0.44 train_steps=1.91
ep100 val_rocauc=0.84485 val_adapt_rocauc=0.89378 adapt_steps=2.02 halt=0.43 train_steps=1.93
[ogbg-molclintox_appnp_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=60 val={'rocauc': 0.9156406973308382} test={'rocauc': 0.7779932575678588} adaptive={'rocauc': 0.886122406422688} steps=2.1621621621621623
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molclintox_appnp_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
|
