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-moltox21 view=gin compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-moltox21 --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=0.73518 val_adapt_rocauc=0.73087 adapt_steps=4.49 halt=0.28 train_steps=2.97
ep20 val_rocauc=0.71039 val_adapt_rocauc=0.70931 adapt_steps=4.19 halt=0.28 train_steps=3.12
ep30 val_rocauc=0.64736 val_adapt_rocauc=0.64143 adapt_steps=2.97 halt=0.28 train_steps=3.01
ep40 val_rocauc=0.72656 val_adapt_rocauc=0.72350 adapt_steps=4.94 halt=0.27 train_steps=3.15
ep50 val_rocauc=0.68867 val_adapt_rocauc=0.72256 adapt_steps=3.45 halt=0.29 train_steps=3.01
ep60 val_rocauc=0.73756 val_adapt_rocauc=0.73596 adapt_steps=5.25 halt=0.28 train_steps=3.07
ep70 val_rocauc=0.76341 val_adapt_rocauc=0.76107 adapt_steps=4.26 halt=0.29 train_steps=3.08
ep80 val_rocauc=0.75486 val_adapt_rocauc=0.76475 adapt_steps=4.17 halt=0.30 train_steps=2.91
ep90 val_rocauc=0.75145 val_adapt_rocauc=0.76691 adapt_steps=3.96 halt=0.31 train_steps=2.86
ep100 val_rocauc=0.75282 val_adapt_rocauc=0.76451 adapt_steps=4.03 halt=0.30 train_steps=2.99
[ogbg-moltox21_gin_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=70 val={'rocauc': 0.763406691780224} test={'rocauc': 0.7044116432874441} adaptive={'rocauc': 0.7111955163056712} steps=4.165816326530612
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-moltox21_gin_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-moltox21 view=gine compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-moltox21 --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=0.66186 val_adapt_rocauc=0.66299 adapt_steps=3.61 halt=0.28 train_steps=3.04
ep20 val_rocauc=0.71740 val_adapt_rocauc=0.70641 adapt_steps=4.29 halt=0.30 train_steps=2.97
ep30 val_rocauc=0.67926 val_adapt_rocauc=0.66414 adapt_steps=3.89 halt=0.28 train_steps=3.03
ep40 val_rocauc=0.71787 val_adapt_rocauc=0.71626 adapt_steps=3.31 halt=0.29 train_steps=2.99
ep50 val_rocauc=0.71826 val_adapt_rocauc=0.72383 adapt_steps=4.48 halt=0.29 train_steps=2.99
ep60 val_rocauc=0.74768 val_adapt_rocauc=0.74921 adapt_steps=4.13 halt=0.29 train_steps=2.98
ep70 val_rocauc=0.75742 val_adapt_rocauc=0.75384 adapt_steps=3.98 halt=0.29 train_steps=3.08
ep80 val_rocauc=0.75643 val_adapt_rocauc=0.74987 adapt_steps=4.43 halt=0.31 train_steps=2.96
ep90 val_rocauc=0.75761 val_adapt_rocauc=0.75256 adapt_steps=4.07 halt=0.30 train_steps=2.99
ep100 val_rocauc=0.76165 val_adapt_rocauc=0.75534 adapt_steps=4.04 halt=0.31 train_steps=2.91
[ogbg-moltox21_gine_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=100 val={'rocauc': 0.7616493170582249} test={'rocauc': 0.7120140634942237} adaptive={'rocauc': 0.7177387150880706} steps=3.931122448979592
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-moltox21_gine_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-moltox21 view=gcn compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-moltox21 --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=0.71516 val_adapt_rocauc=0.71843 adapt_steps=2.53 halt=0.29 train_steps=2.97
ep20 val_rocauc=0.72992 val_adapt_rocauc=0.72176 adapt_steps=3.40 halt=0.30 train_steps=2.91
ep30 val_rocauc=0.71748 val_adapt_rocauc=0.73422 adapt_steps=2.90 halt=0.30 train_steps=2.89
ep40 val_rocauc=0.75509 val_adapt_rocauc=0.75981 adapt_steps=2.73 halt=0.32 train_steps=2.77
ep50 val_rocauc=0.75894 val_adapt_rocauc=0.75514 adapt_steps=3.85 halt=0.33 train_steps=2.68
ep60 val_rocauc=0.76775 val_adapt_rocauc=0.77369 adapt_steps=3.11 halt=0.35 train_steps=2.52
ep70 val_rocauc=0.75412 val_adapt_rocauc=0.76445 adapt_steps=2.73 halt=0.38 train_steps=2.25
ep80 val_rocauc=0.74793 val_adapt_rocauc=0.76557 adapt_steps=2.42 halt=0.40 train_steps=2.14
ep90 val_rocauc=0.74992 val_adapt_rocauc=0.76933 adapt_steps=2.46 halt=0.42 train_steps=2.04
ep100 val_rocauc=0.74961 val_adapt_rocauc=0.76962 adapt_steps=2.47 halt=0.42 train_steps=1.98
[ogbg-moltox21_gcn_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=60 val={'rocauc': 0.7677478126932855} test={'rocauc': 0.7135675459281269} adaptive={'rocauc': 0.7247326710130202} steps=3.2589285714285716
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-moltox21_gcn_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-moltox21 view=graphsage compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-moltox21 --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=0.66353 val_adapt_rocauc=0.66241 adapt_steps=6.17 halt=0.30 train_steps=2.79
ep20 val_rocauc=0.69687 val_adapt_rocauc=0.71600 adapt_steps=3.03 halt=0.28 train_steps=3.04
ep30 val_rocauc=0.74011 val_adapt_rocauc=0.73276 adapt_steps=3.99 halt=0.30 train_steps=2.91
ep40 val_rocauc=0.77334 val_adapt_rocauc=0.77752 adapt_steps=4.25 halt=0.31 train_steps=2.82
ep50 val_rocauc=0.77022 val_adapt_rocauc=0.78137 adapt_steps=3.56 halt=0.32 train_steps=2.78
ep60 val_rocauc=0.77918 val_adapt_rocauc=0.77520 adapt_steps=3.29 halt=0.34 train_steps=2.65
ep70 val_rocauc=0.77387 val_adapt_rocauc=0.77500 adapt_steps=2.79 halt=0.37 train_steps=2.37
ep80 val_rocauc=0.76848 val_adapt_rocauc=0.77284 adapt_steps=2.96 halt=0.40 train_steps=2.15
ep90 val_rocauc=0.76947 val_adapt_rocauc=0.77447 adapt_steps=2.75 halt=0.40 train_steps=2.14
ep100 val_rocauc=0.77243 val_adapt_rocauc=0.77696 adapt_steps=2.73 halt=0.40 train_steps=2.18
[ogbg-moltox21_graphsage_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=60 val={'rocauc': 0.7791810357079593} test={'rocauc': 0.7217088253529022} adaptive={'rocauc': 0.7282454305579198} steps=3.3903061224489797
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-moltox21_graphsage_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-moltox21 view=gatv2 compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-moltox21 --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=0.68214 val_adapt_rocauc=0.66532 adapt_steps=3.94 halt=0.27 train_steps=3.13
ep20 val_rocauc=0.74877 val_adapt_rocauc=0.73707 adapt_steps=4.38 halt=0.29 train_steps=2.96
ep30 val_rocauc=0.76735 val_adapt_rocauc=0.76425 adapt_steps=4.05 halt=0.30 train_steps=2.88
ep40 val_rocauc=0.74010 val_adapt_rocauc=0.74347 adapt_steps=3.25 halt=0.31 train_steps=2.84
ep50 val_rocauc=0.75964 val_adapt_rocauc=0.75650 adapt_steps=4.19 halt=0.33 train_steps=2.73
ep60 val_rocauc=0.76484 val_adapt_rocauc=0.76741 adapt_steps=3.80 halt=0.30 train_steps=2.94
ep70 val_rocauc=0.75529 val_adapt_rocauc=0.76640 adapt_steps=3.48 halt=0.35 train_steps=2.56
ep80 val_rocauc=0.74656 val_adapt_rocauc=0.75879 adapt_steps=2.90 halt=0.36 train_steps=2.46
ep90 val_rocauc=0.73882 val_adapt_rocauc=0.75668 adapt_steps=3.01 halt=0.39 train_steps=2.26
ep100 val_rocauc=0.73632 val_adapt_rocauc=0.75539 adapt_steps=2.97 halt=0.38 train_steps=2.35
[ogbg-moltox21_gatv2_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=30 val={'rocauc': 0.767354951757618} test={'rocauc': 0.7155377552729235} adaptive={'rocauc': 0.7172144629316312} steps=4.042091836734694
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-moltox21_gatv2_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-moltox21 view=graphconv compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-moltox21 --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=0.70226 val_adapt_rocauc=0.68368 adapt_steps=4.53 halt=0.28 train_steps=3.06
ep20 val_rocauc=0.69482 val_adapt_rocauc=0.69973 adapt_steps=2.94 halt=0.28 train_steps=3.02
ep30 val_rocauc=0.73997 val_adapt_rocauc=0.72387 adapt_steps=3.56 halt=0.29 train_steps=2.95
ep40 val_rocauc=0.73715 val_adapt_rocauc=0.74654 adapt_steps=3.93 halt=0.30 train_steps=2.93
ep50 val_rocauc=0.76406 val_adapt_rocauc=0.76634 adapt_steps=4.12 halt=0.33 train_steps=2.73
ep60 val_rocauc=0.74500 val_adapt_rocauc=0.77410 adapt_steps=3.28 halt=0.35 train_steps=2.53
ep70 val_rocauc=0.74134 val_adapt_rocauc=0.76240 adapt_steps=3.34 halt=0.38 train_steps=2.30
ep80 val_rocauc=0.74270 val_adapt_rocauc=0.76980 adapt_steps=3.36 halt=0.37 train_steps=2.39
ep90 val_rocauc=0.74651 val_adapt_rocauc=0.76905 adapt_steps=2.91 halt=0.39 train_steps=2.26
ep100 val_rocauc=0.74268 val_adapt_rocauc=0.76713 adapt_steps=2.77 halt=0.38 train_steps=2.29
[ogbg-moltox21_graphconv_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=50 val={'rocauc': 0.7640586359961631} test={'rocauc': 0.721040156330531} adaptive={'rocauc': 0.732265876016213} steps=4.089285714285714
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-moltox21_graphconv_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-moltox21 view=transformer compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-moltox21 --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=0.67594 val_adapt_rocauc=0.67877 adapt_steps=2.90 halt=0.27 train_steps=3.04
ep20 val_rocauc=0.74853 val_adapt_rocauc=0.75119 adapt_steps=4.17 halt=0.28 train_steps=3.04
ep30 val_rocauc=0.75978 val_adapt_rocauc=0.77013 adapt_steps=3.66 halt=0.28 train_steps=3.06
ep40 val_rocauc=0.78399 val_adapt_rocauc=0.79071 adapt_steps=4.20 halt=0.32 train_steps=2.76
ep50 val_rocauc=0.76021 val_adapt_rocauc=0.77394 adapt_steps=3.78 halt=0.32 train_steps=2.70
ep60 val_rocauc=0.77603 val_adapt_rocauc=0.77979 adapt_steps=3.46 halt=0.35 train_steps=2.50
ep70 val_rocauc=0.78021 val_adapt_rocauc=0.78970 adapt_steps=2.86 halt=0.39 train_steps=2.27
ep80 val_rocauc=0.77594 val_adapt_rocauc=0.78551 adapt_steps=2.57 halt=0.42 train_steps=2.00
ep90 val_rocauc=0.77827 val_adapt_rocauc=0.78340 adapt_steps=2.53 halt=0.42 train_steps=2.00
ep100 val_rocauc=0.77925 val_adapt_rocauc=0.78691 adapt_steps=2.50 halt=0.43 train_steps=1.93
[ogbg-moltox21_transformer_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=40 val={'rocauc': 0.7839946999002233} test={'rocauc': 0.723701343716436} adaptive={'rocauc': 0.724166086603444} steps=4.32780612244898
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-moltox21_transformer_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-moltox21 view=pna compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-moltox21 --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=0.66669 val_adapt_rocauc=0.66243 adapt_steps=7.30 halt=0.29 train_steps=3.00
/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=0.69019 val_adapt_rocauc=0.67918 adapt_steps=5.84 halt=0.30 train_steps=2.85
/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=0.70422 val_adapt_rocauc=0.69523 adapt_steps=4.71 halt=0.28 train_steps=3.13
/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=0.71722 val_adapt_rocauc=0.71965 adapt_steps=2.52 halt=0.29 train_steps=2.95
/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=0.70706 val_adapt_rocauc=0.70996 adapt_steps=6.01 halt=0.30 train_steps=2.87
/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=0.73991 val_adapt_rocauc=0.74273 adapt_steps=4.85 halt=0.29 train_steps=2.98
/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=0.72423 val_adapt_rocauc=0.73134 adapt_steps=2.85 halt=0.31 train_steps=2.81
/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=0.75108 val_adapt_rocauc=0.74780 adapt_steps=4.23 halt=0.30 train_steps=2.91
/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(
ep90 val_rocauc=0.76247 val_adapt_rocauc=0.75804 adapt_steps=3.94 halt=0.32 train_steps=2.79
/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=0.75859 val_adapt_rocauc=0.75525 adapt_steps=3.88 halt=0.31 train_steps=2.83
[ogbg-moltox21_pna_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=90 val={'rocauc': 0.7624732674125697} test={'rocauc': 0.7137965144412185} adaptive={'rocauc': 0.7200143809038219} steps=4.058673469387755
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-moltox21_pna_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-moltox21 view=gen compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-moltox21 --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=0.65402 val_adapt_rocauc=0.65164 adapt_steps=3.07 halt=0.28 train_steps=2.91
ep20 val_rocauc=0.67375 val_adapt_rocauc=0.66200 adapt_steps=3.13 halt=0.27 train_steps=3.10
ep30 val_rocauc=0.70573 val_adapt_rocauc=0.70756 adapt_steps=3.81 halt=0.28 train_steps=3.02
ep40 val_rocauc=0.73605 val_adapt_rocauc=0.74012 adapt_steps=3.52 halt=0.27 train_steps=3.16
ep50 val_rocauc=0.71987 val_adapt_rocauc=0.73228 adapt_steps=3.73 halt=0.28 train_steps=3.07
ep60 val_rocauc=0.74139 val_adapt_rocauc=0.73571 adapt_steps=3.62 halt=0.29 train_steps=3.01
ep70 val_rocauc=0.74691 val_adapt_rocauc=0.74235 adapt_steps=3.40 halt=0.29 train_steps=3.05
ep80 val_rocauc=0.75480 val_adapt_rocauc=0.76023 adapt_steps=3.50 halt=0.30 train_steps=2.97
ep90 val_rocauc=0.75153 val_adapt_rocauc=0.75746 adapt_steps=3.76 halt=0.30 train_steps=2.96
ep100 val_rocauc=0.75081 val_adapt_rocauc=0.76030 adapt_steps=3.70 halt=0.31 train_steps=2.86
[ogbg-moltox21_gen_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=80 val={'rocauc': 0.7548045767051833} test={'rocauc': 0.7300226043604717} adaptive={'rocauc': 0.7365872982320871} steps=3.433673469387755
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-moltox21_gen_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-moltox21 view=film compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-moltox21 --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=0.57757 val_adapt_rocauc=0.58541 adapt_steps=3.29 halt=0.31 train_steps=2.50
ep20 val_rocauc=0.55966 val_adapt_rocauc=0.57096 adapt_steps=4.07 halt=0.28 train_steps=2.95
ep30 val_rocauc=0.67643 val_adapt_rocauc=0.69130 adapt_steps=3.16 halt=0.28 train_steps=3.07
ep40 val_rocauc=0.69186 val_adapt_rocauc=0.69814 adapt_steps=4.81 halt=0.28 train_steps=3.08
ep50 val_rocauc=0.65619 val_adapt_rocauc=0.67613 adapt_steps=3.51 halt=0.27 train_steps=3.19
ep60 val_rocauc=0.70395 val_adapt_rocauc=0.71660 adapt_steps=3.19 halt=0.31 train_steps=2.93
ep70 val_rocauc=0.72848 val_adapt_rocauc=0.73235 adapt_steps=3.83 halt=0.31 train_steps=2.87
ep80 val_rocauc=0.74847 val_adapt_rocauc=0.75433 adapt_steps=4.28 halt=0.30 train_steps=3.00
ep90 val_rocauc=0.74854 val_adapt_rocauc=0.75584 adapt_steps=3.98 halt=0.32 train_steps=2.89
ep100 val_rocauc=0.75129 val_adapt_rocauc=0.75550 adapt_steps=3.77 halt=0.32 train_steps=2.83
[ogbg-moltox21_film_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=100 val={'rocauc': 0.7512888457712221} test={'rocauc': 0.6885051838242511} adaptive={'rocauc': 0.6897262581951776} steps=3.55484693877551
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-moltox21_film_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-moltox21 view=resgated compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-moltox21 --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=0.69369 val_adapt_rocauc=0.68651 adapt_steps=3.85 halt=0.28 train_steps=3.03
ep20 val_rocauc=0.73254 val_adapt_rocauc=0.71275 adapt_steps=5.04 halt=0.27 train_steps=3.19
ep30 val_rocauc=0.70299 val_adapt_rocauc=0.70576 adapt_steps=3.07 halt=0.31 train_steps=2.80
ep40 val_rocauc=0.76163 val_adapt_rocauc=0.76283 adapt_steps=4.54 halt=0.30 train_steps=2.90
ep50 val_rocauc=0.75392 val_adapt_rocauc=0.75554 adapt_steps=4.28 halt=0.32 train_steps=2.71
ep60 val_rocauc=0.77327 val_adapt_rocauc=0.77823 adapt_steps=3.05 halt=0.35 train_steps=2.49
ep70 val_rocauc=0.77098 val_adapt_rocauc=0.77495 adapt_steps=3.01 halt=0.37 train_steps=2.37
ep80 val_rocauc=0.77372 val_adapt_rocauc=0.78024 adapt_steps=2.81 halt=0.39 train_steps=2.26
ep90 val_rocauc=0.77062 val_adapt_rocauc=0.77479 adapt_steps=2.66 halt=0.41 train_steps=2.11
ep100 val_rocauc=0.76956 val_adapt_rocauc=0.77366 adapt_steps=2.72 halt=0.41 train_steps=2.10
[ogbg-moltox21_resgated_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=80 val={'rocauc': 0.7737227459288354} test={'rocauc': 0.7454181481867773} adaptive={'rocauc': 0.7448724120402496} steps=2.7767857142857144
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-moltox21_resgated_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-moltox21 view=tag compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-moltox21 --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=0.65152 val_adapt_rocauc=0.63598 adapt_steps=5.82 halt=0.28 train_steps=3.00
ep20 val_rocauc=0.71815 val_adapt_rocauc=0.71620 adapt_steps=3.05 halt=0.28 train_steps=3.03
ep30 val_rocauc=0.72675 val_adapt_rocauc=0.72953 adapt_steps=5.25 halt=0.30 train_steps=2.85
ep40 val_rocauc=0.70713 val_adapt_rocauc=0.71988 adapt_steps=4.60 halt=0.30 train_steps=2.87
ep50 val_rocauc=0.73557 val_adapt_rocauc=0.74105 adapt_steps=3.83 halt=0.31 train_steps=2.81
ep60 val_rocauc=0.75468 val_adapt_rocauc=0.76648 adapt_steps=3.73 halt=0.31 train_steps=2.88
ep70 val_rocauc=0.77518 val_adapt_rocauc=0.77338 adapt_steps=3.67 halt=0.33 train_steps=2.67
ep80 val_rocauc=0.77918 val_adapt_rocauc=0.78370 adapt_steps=3.28 halt=0.35 train_steps=2.52
ep90 val_rocauc=0.77746 val_adapt_rocauc=0.77301 adapt_steps=3.44 halt=0.36 train_steps=2.46
ep100 val_rocauc=0.77192 val_adapt_rocauc=0.76945 adapt_steps=3.25 halt=0.36 train_steps=2.43
[ogbg-moltox21_tag_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=80 val={'rocauc': 0.7791811026208421} test={'rocauc': 0.7194750067722034} adaptive={'rocauc': 0.7225403729796881} steps=3.318877551020408
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-moltox21_tag_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-moltox21 view=sgc compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-moltox21 --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=0.58511 val_adapt_rocauc=0.58665 adapt_steps=2.96 halt=0.29 train_steps=2.97
ep20 val_rocauc=0.72203 val_adapt_rocauc=0.68935 adapt_steps=3.69 halt=0.29 train_steps=2.92
ep30 val_rocauc=0.74058 val_adapt_rocauc=0.74146 adapt_steps=4.68 halt=0.28 train_steps=3.06
ep40 val_rocauc=0.74632 val_adapt_rocauc=0.74580 adapt_steps=4.82 halt=0.29 train_steps=3.01
ep50 val_rocauc=0.72977 val_adapt_rocauc=0.73373 adapt_steps=3.75 halt=0.31 train_steps=2.89
ep60 val_rocauc=0.74202 val_adapt_rocauc=0.74441 adapt_steps=3.87 halt=0.32 train_steps=2.77
ep70 val_rocauc=0.76058 val_adapt_rocauc=0.75527 adapt_steps=3.41 halt=0.34 train_steps=2.64
ep80 val_rocauc=0.76323 val_adapt_rocauc=0.76205 adapt_steps=3.39 halt=0.36 train_steps=2.47
ep90 val_rocauc=0.76418 val_adapt_rocauc=0.76382 adapt_steps=3.20 halt=0.37 train_steps=2.39
ep100 val_rocauc=0.76686 val_adapt_rocauc=0.76364 adapt_steps=3.21 halt=0.36 train_steps=2.44
[ogbg-moltox21_sgc_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=100 val={'rocauc': 0.7668620788836082} test={'rocauc': 0.7261733029463184} adaptive={'rocauc': 0.7330710041428746} steps=3.260204081632653
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-moltox21_sgc_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-moltox21 view=cheb compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-moltox21 --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=0.71726 val_adapt_rocauc=0.71158 adapt_steps=2.69 halt=0.30 train_steps=2.93
ep20 val_rocauc=0.74514 val_adapt_rocauc=0.74244 adapt_steps=3.13 halt=0.29 train_steps=2.94
ep30 val_rocauc=0.74745 val_adapt_rocauc=0.74514 adapt_steps=4.27 halt=0.31 train_steps=2.82
ep40 val_rocauc=0.75855 val_adapt_rocauc=0.75693 adapt_steps=3.17 halt=0.32 train_steps=2.71
ep50 val_rocauc=0.77450 val_adapt_rocauc=0.77541 adapt_steps=3.24 halt=0.35 train_steps=2.52
ep60 val_rocauc=0.78240 val_adapt_rocauc=0.78798 adapt_steps=2.92 halt=0.38 train_steps=2.30
ep70 val_rocauc=0.77174 val_adapt_rocauc=0.78493 adapt_steps=2.44 halt=0.42 train_steps=1.98
ep80 val_rocauc=0.76722 val_adapt_rocauc=0.78035 adapt_steps=2.14 halt=0.45 train_steps=1.75
ep90 val_rocauc=0.76854 val_adapt_rocauc=0.78215 adapt_steps=2.12 halt=0.47 train_steps=1.67
ep100 val_rocauc=0.77046 val_adapt_rocauc=0.78353 adapt_steps=2.12 halt=0.46 train_steps=1.67
[ogbg-moltox21_cheb_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=60 val={'rocauc': 0.7823963460925124} test={'rocauc': 0.7195427399764524} adaptive={'rocauc': 0.7202976536434207} steps=3.014030612244898
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-moltox21_cheb_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-moltox21 view=arma compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-moltox21 --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=0.63590 val_adapt_rocauc=0.63888 adapt_steps=2.72 halt=0.29 train_steps=2.91
ep20 val_rocauc=0.71067 val_adapt_rocauc=0.70324 adapt_steps=4.82 halt=0.30 train_steps=2.86
ep30 val_rocauc=0.75699 val_adapt_rocauc=0.76269 adapt_steps=4.94 halt=0.29 train_steps=2.96
ep40 val_rocauc=0.75283 val_adapt_rocauc=0.76066 adapt_steps=2.89 halt=0.31 train_steps=2.87
ep50 val_rocauc=0.74882 val_adapt_rocauc=0.75319 adapt_steps=2.96 halt=0.32 train_steps=2.76
ep60 val_rocauc=0.74867 val_adapt_rocauc=0.75550 adapt_steps=2.58 halt=0.34 train_steps=2.61
ep70 val_rocauc=0.75749 val_adapt_rocauc=0.77569 adapt_steps=2.86 halt=0.38 train_steps=2.31
ep80 val_rocauc=0.73534 val_adapt_rocauc=0.77063 adapt_steps=2.44 halt=0.40 train_steps=2.10
ep90 val_rocauc=0.74163 val_adapt_rocauc=0.77322 adapt_steps=2.37 halt=0.41 train_steps=2.04
ep100 val_rocauc=0.74596 val_adapt_rocauc=0.77377 adapt_steps=2.39 halt=0.42 train_steps=1.95
[ogbg-moltox21_arma_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=70 val={'rocauc': 0.7574890736647663} test={'rocauc': 0.7161943072626955} adaptive={'rocauc': 0.7237147258339075} steps=2.868622448979592
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-moltox21_arma_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-moltox21 view=mf compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-moltox21 --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=0.68103 val_adapt_rocauc=0.67327 adapt_steps=6.68 halt=0.28 train_steps=3.06
ep20 val_rocauc=0.71526 val_adapt_rocauc=0.70565 adapt_steps=3.58 halt=0.27 train_steps=3.17
ep30 val_rocauc=0.73936 val_adapt_rocauc=0.73882 adapt_steps=4.88 halt=0.30 train_steps=2.92
ep40 val_rocauc=0.68475 val_adapt_rocauc=0.69545 adapt_steps=3.09 halt=0.32 train_steps=2.71
ep50 val_rocauc=0.72737 val_adapt_rocauc=0.73189 adapt_steps=3.75 halt=0.33 train_steps=2.71
ep60 val_rocauc=0.76486 val_adapt_rocauc=0.77034 adapt_steps=3.49 halt=0.35 train_steps=2.53
ep70 val_rocauc=0.76555 val_adapt_rocauc=0.77366 adapt_steps=2.97 halt=0.37 train_steps=2.39
ep80 val_rocauc=0.76124 val_adapt_rocauc=0.77200 adapt_steps=2.91 halt=0.40 train_steps=2.18
ep90 val_rocauc=0.75552 val_adapt_rocauc=0.76936 adapt_steps=2.66 halt=0.40 train_steps=2.13
ep100 val_rocauc=0.75477 val_adapt_rocauc=0.76857 adapt_steps=2.64 halt=0.42 train_steps=2.02
[ogbg-moltox21_mf_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=70 val={'rocauc': 0.76555289309204} test={'rocauc': 0.7235855082035453} adaptive={'rocauc': 0.7314152133751847} steps=2.9566326530612246
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-moltox21_mf_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-moltox21 view=appnp compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:1
python3 rrog/train_ogb_graphprop.py --dataset ogbg-moltox21 --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=0.66956 val_adapt_rocauc=0.68596 adapt_steps=2.92 halt=0.30 train_steps=2.86
ep20 val_rocauc=0.66741 val_adapt_rocauc=0.68976 adapt_steps=2.90 halt=0.31 train_steps=2.80
ep30 val_rocauc=0.69725 val_adapt_rocauc=0.69142 adapt_steps=3.11 halt=0.33 train_steps=2.66
ep40 val_rocauc=0.70858 val_adapt_rocauc=0.72396 adapt_steps=3.15 halt=0.36 train_steps=2.47
ep50 val_rocauc=0.71714 val_adapt_rocauc=0.73276 adapt_steps=2.59 halt=0.39 train_steps=2.24
ep60 val_rocauc=0.70738 val_adapt_rocauc=0.72373 adapt_steps=2.53 halt=0.42 train_steps=1.95
ep70 val_rocauc=0.69595 val_adapt_rocauc=0.71924 adapt_steps=2.25 halt=0.43 train_steps=1.93
ep80 val_rocauc=0.70875 val_adapt_rocauc=0.72985 adapt_steps=2.16 halt=0.44 train_steps=1.86
ep90 val_rocauc=0.70758 val_adapt_rocauc=0.72805 adapt_steps=2.16 halt=0.44 train_steps=1.82
ep100 val_rocauc=0.70806 val_adapt_rocauc=0.72668 adapt_steps=2.16 halt=0.44 train_steps=1.83
[ogbg-moltox21_appnp_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=50 val={'rocauc': 0.7171412864790295} test={'rocauc': 0.7196990667216929} adaptive={'rocauc': 0.7331958989125181} steps=2.586734693877551
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-moltox21_appnp_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
|
