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-molbace view=gin compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbace --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:0 --num_workers 0
ep10 val_rocauc=0.62747 val_adapt_rocauc=0.63956 adapt_steps=5.64 halt=0.17 train_steps=4.09
ep20 val_rocauc=0.40916 val_adapt_rocauc=0.38755 adapt_steps=2.62 halt=0.15 train_steps=4.15
ep30 val_rocauc=0.45714 val_adapt_rocauc=0.49011 adapt_steps=5.73 halt=0.17 train_steps=4.04
ep40 val_rocauc=0.56740 val_adapt_rocauc=0.60037 adapt_steps=3.81 halt=0.24 train_steps=3.51
ep50 val_rocauc=0.43443 val_adapt_rocauc=0.43736 adapt_steps=4.63 halt=0.20 train_steps=3.57
ep60 val_rocauc=0.66557 val_adapt_rocauc=0.67289 adapt_steps=4.50 halt=0.22 train_steps=3.60
ep70 val_rocauc=0.65018 val_adapt_rocauc=0.65604 adapt_steps=4.30 halt=0.27 train_steps=3.14
ep80 val_rocauc=0.68022 val_adapt_rocauc=0.66520 adapt_steps=4.40 halt=0.30 train_steps=2.85
ep90 val_rocauc=0.67216 val_adapt_rocauc=0.67070 adapt_steps=4.48 halt=0.27 train_steps=3.13
ep100 val_rocauc=0.67582 val_adapt_rocauc=0.66447 adapt_steps=4.34 halt=0.30 train_steps=2.93
[ogbg-molbace_gin_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=80 val={'rocauc': 0.6802197802197802} test={'rocauc': 0.723526343244653} adaptive={'rocauc': 0.721961398017736} steps=3.473684210526316
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbace_gin_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molbace view=gine compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbace --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:0 --num_workers 0
ep10 val_rocauc=0.58205 val_adapt_rocauc=0.58462 adapt_steps=7.00 halt=0.27 train_steps=3.01
ep20 val_rocauc=0.54835 val_adapt_rocauc=0.55495 adapt_steps=7.38 halt=0.13 train_steps=4.38
ep30 val_rocauc=0.64542 val_adapt_rocauc=0.57033 adapt_steps=4.30 halt=0.15 train_steps=4.25
ep40 val_rocauc=0.47033 val_adapt_rocauc=0.49597 adapt_steps=4.36 halt=0.18 train_steps=3.89
ep50 val_rocauc=0.57619 val_adapt_rocauc=0.55971 adapt_steps=4.50 halt=0.18 train_steps=3.77
ep60 val_rocauc=0.56593 val_adapt_rocauc=0.57143 adapt_steps=4.76 halt=0.24 train_steps=3.54
ep70 val_rocauc=0.68242 val_adapt_rocauc=0.69780 adapt_steps=3.98 halt=0.25 train_steps=3.42
ep80 val_rocauc=0.65897 val_adapt_rocauc=0.66081 adapt_steps=4.22 halt=0.26 train_steps=3.37
ep90 val_rocauc=0.63187 val_adapt_rocauc=0.63150 adapt_steps=4.58 halt=0.24 train_steps=3.38
ep100 val_rocauc=0.62381 val_adapt_rocauc=0.62637 adapt_steps=4.89 halt=0.25 train_steps=3.32
[ogbg-molbace_gine_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=70 val={'rocauc': 0.6824175824175824} test={'rocauc': 0.7718657624760911} adaptive={'rocauc': 0.7697791688402017} steps=3.1710526315789473
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbace_gine_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molbace view=gcn compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbace --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:0 --num_workers 0
ep10 val_rocauc=0.59927 val_adapt_rocauc=0.60696 adapt_steps=5.45 halt=0.21 train_steps=3.53
ep20 val_rocauc=0.61245 val_adapt_rocauc=0.60147 adapt_steps=4.89 halt=0.21 train_steps=3.54
ep30 val_rocauc=0.71941 val_adapt_rocauc=0.72015 adapt_steps=4.09 halt=0.24 train_steps=3.64
ep40 val_rocauc=0.66923 val_adapt_rocauc=0.67692 adapt_steps=4.70 halt=0.21 train_steps=3.73
ep50 val_rocauc=0.72015 val_adapt_rocauc=0.70220 adapt_steps=4.28 halt=0.27 train_steps=3.33
ep60 val_rocauc=0.66264 val_adapt_rocauc=0.67546 adapt_steps=4.85 halt=0.30 train_steps=2.94
ep70 val_rocauc=0.57619 val_adapt_rocauc=0.57766 adapt_steps=3.40 halt=0.29 train_steps=3.02
ep80 val_rocauc=0.60147 val_adapt_rocauc=0.61648 adapt_steps=3.09 halt=0.34 train_steps=2.52
ep90 val_rocauc=0.63480 val_adapt_rocauc=0.63773 adapt_steps=3.01 halt=0.37 train_steps=2.41
ep100 val_rocauc=0.63810 val_adapt_rocauc=0.63883 adapt_steps=2.89 halt=0.39 train_steps=2.20
[ogbg-molbace_gcn_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=50 val={'rocauc': 0.7201465201465201} test={'rocauc': 0.7483915840723353} adaptive={'rocauc': 0.7450878108155105} steps=3.960526315789474
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbace_gcn_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molbace view=graphsage compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbace --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:0 --num_workers 0
ep10 val_rocauc=0.57070 val_adapt_rocauc=0.60293 adapt_steps=5.37 halt=0.30 train_steps=2.89
ep20 val_rocauc=0.56520 val_adapt_rocauc=0.57692 adapt_steps=6.33 halt=0.20 train_steps=3.67
ep30 val_rocauc=0.59670 val_adapt_rocauc=0.60916 adapt_steps=3.97 halt=0.25 train_steps=3.37
ep40 val_rocauc=0.62564 val_adapt_rocauc=0.62454 adapt_steps=2.83 halt=0.30 train_steps=3.03
ep50 val_rocauc=0.69634 val_adapt_rocauc=0.69597 adapt_steps=4.89 halt=0.23 train_steps=3.60
ep60 val_rocauc=0.64286 val_adapt_rocauc=0.63480 adapt_steps=3.69 halt=0.26 train_steps=3.31
ep70 val_rocauc=0.70916 val_adapt_rocauc=0.71502 adapt_steps=4.54 halt=0.26 train_steps=3.45
ep80 val_rocauc=0.63626 val_adapt_rocauc=0.65897 adapt_steps=4.31 halt=0.28 train_steps=3.07
ep90 val_rocauc=0.64066 val_adapt_rocauc=0.63590 adapt_steps=3.86 halt=0.32 train_steps=2.74
ep100 val_rocauc=0.64066 val_adapt_rocauc=0.64835 adapt_steps=3.68 halt=0.30 train_steps=2.89
[ogbg-molbace_graphsage_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=70 val={'rocauc': 0.7091575091575092} test={'rocauc': 0.7821248478525474} adaptive={'rocauc': 0.7909928708050773} steps=3.75
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbace_graphsage_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molbace view=gatv2 compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbace --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:0 --num_workers 0
ep10 val_rocauc=0.68608 val_adapt_rocauc=0.68938 adapt_steps=7.89 halt=0.17 train_steps=3.74
ep20 val_rocauc=0.63260 val_adapt_rocauc=0.62674 adapt_steps=5.79 halt=0.21 train_steps=3.57
ep30 val_rocauc=0.61612 val_adapt_rocauc=0.64615 adapt_steps=5.33 halt=0.23 train_steps=3.07
ep40 val_rocauc=0.71319 val_adapt_rocauc=0.71978 adapt_steps=4.52 halt=0.23 train_steps=3.55
ep50 val_rocauc=0.66960 val_adapt_rocauc=0.65458 adapt_steps=5.34 halt=0.21 train_steps=3.69
ep60 val_rocauc=0.69927 val_adapt_rocauc=0.68571 adapt_steps=5.17 halt=0.21 train_steps=3.63
ep70 val_rocauc=0.63956 val_adapt_rocauc=0.64652 adapt_steps=4.17 halt=0.28 train_steps=3.24
ep80 val_rocauc=0.63333 val_adapt_rocauc=0.63260 adapt_steps=3.86 halt=0.26 train_steps=3.23
ep90 val_rocauc=0.64103 val_adapt_rocauc=0.65714 adapt_steps=2.97 halt=0.34 train_steps=2.71
ep100 val_rocauc=0.63297 val_adapt_rocauc=0.63773 adapt_steps=3.26 halt=0.32 train_steps=2.75
[ogbg-molbace_gatv2_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=40 val={'rocauc': 0.7131868131868132} test={'rocauc': 0.8261171970092158} adaptive={'rocauc': 0.8262910798122065} steps=4.006578947368421
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbace_gatv2_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molbace view=graphconv compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbace --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:0 --num_workers 0
ep10 val_rocauc=0.56081 val_adapt_rocauc=0.58388 adapt_steps=6.43 halt=0.27 train_steps=3.07
ep20 val_rocauc=0.61245 val_adapt_rocauc=0.59780 adapt_steps=7.42 halt=0.18 train_steps=3.93
ep30 val_rocauc=0.69084 val_adapt_rocauc=0.69231 adapt_steps=5.89 halt=0.22 train_steps=3.39
ep40 val_rocauc=0.66154 val_adapt_rocauc=0.66520 adapt_steps=3.93 halt=0.21 train_steps=3.61
ep50 val_rocauc=0.57399 val_adapt_rocauc=0.60147 adapt_steps=3.54 halt=0.32 train_steps=2.79
ep60 val_rocauc=0.67546 val_adapt_rocauc=0.69011 adapt_steps=3.45 halt=0.30 train_steps=2.92
ep70 val_rocauc=0.61099 val_adapt_rocauc=0.62564 adapt_steps=3.86 halt=0.33 train_steps=2.69
ep80 val_rocauc=0.65275 val_adapt_rocauc=0.65458 adapt_steps=3.17 halt=0.37 train_steps=2.41
ep90 val_rocauc=0.66740 val_adapt_rocauc=0.66703 adapt_steps=3.09 halt=0.40 train_steps=2.22
ep100 val_rocauc=0.67766 val_adapt_rocauc=0.67875 adapt_steps=3.11 halt=0.36 train_steps=2.30
[ogbg-molbace_graphconv_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=30 val={'rocauc': 0.6908424908424909} test={'rocauc': 0.7475221700573813} adaptive={'rocauc': 0.7638671535385151} steps=4.276315789473684
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbace_graphconv_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molbace view=transformer compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbace --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:0 --num_workers 0
ep10 val_rocauc=0.63223 val_adapt_rocauc=0.64322 adapt_steps=6.61 halt=0.18 train_steps=3.96
ep20 val_rocauc=0.54066 val_adapt_rocauc=0.57179 adapt_steps=7.52 halt=0.18 train_steps=3.90
ep30 val_rocauc=0.51538 val_adapt_rocauc=0.54872 adapt_steps=3.85 halt=0.19 train_steps=3.74
ep40 val_rocauc=0.61978 val_adapt_rocauc=0.59597 adapt_steps=3.91 halt=0.24 train_steps=3.50
ep50 val_rocauc=0.57106 val_adapt_rocauc=0.58095 adapt_steps=3.20 halt=0.26 train_steps=3.27
ep60 val_rocauc=0.62967 val_adapt_rocauc=0.62747 adapt_steps=5.23 halt=0.28 train_steps=3.29
ep70 val_rocauc=0.61575 val_adapt_rocauc=0.62857 adapt_steps=3.82 halt=0.35 train_steps=2.50
ep80 val_rocauc=0.65348 val_adapt_rocauc=0.65385 adapt_steps=3.36 halt=0.37 train_steps=2.42
ep90 val_rocauc=0.64103 val_adapt_rocauc=0.63114 adapt_steps=3.39 halt=0.36 train_steps=2.45
ep100 val_rocauc=0.64396 val_adapt_rocauc=0.62784 adapt_steps=3.23 halt=0.37 train_steps=2.26
[ogbg-molbace_transformer_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=80 val={'rocauc': 0.6534798534798535} test={'rocauc': 0.7191792731698834} adaptive={'rocauc': 0.729264475743349} steps=2.710526315789474
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbace_transformer_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molbace view=pna compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbace --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:0 --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=0.59634 val_adapt_rocauc=0.59707 adapt_steps=7.95 halt=0.19 train_steps=3.73
/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=0.70952 val_adapt_rocauc=0.58462 adapt_steps=3.56 halt=0.17 train_steps=3.89
/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=0.62454 val_adapt_rocauc=0.62454 adapt_steps=7.94 halt=0.18 train_steps=3.65
/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=0.50879 val_adapt_rocauc=0.50623 adapt_steps=7.53 halt=0.13 train_steps=4.40
/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(
ep50 val_rocauc=0.51941 val_adapt_rocauc=0.51319 adapt_steps=6.01 halt=0.18 train_steps=3.86
/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=0.64579 val_adapt_rocauc=0.65238 adapt_steps=6.48 halt=0.18 train_steps=3.94
/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=0.64249 val_adapt_rocauc=0.63333 adapt_steps=4.91 halt=0.24 train_steps=3.31
/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=0.60293 val_adapt_rocauc=0.60806 adapt_steps=5.57 halt=0.22 train_steps=3.46
/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=0.61209 val_adapt_rocauc=0.60952 adapt_steps=5.67 halt=0.23 train_steps=3.43
/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=0.61465 val_adapt_rocauc=0.62491 adapt_steps=6.01 halt=0.23 train_steps=3.48
[ogbg-molbace_pna_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=20 val={'rocauc': 0.7095238095238096} test={'rocauc': 0.7809076682316118} adaptive={'rocauc': 0.7337854286211093} steps=3.914473684210526
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbace_pna_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molbace view=gen compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbace --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:0 --num_workers 0
ep10 val_rocauc=0.43187 val_adapt_rocauc=0.56960 adapt_steps=3.77 halt=0.20 train_steps=4.07
ep20 val_rocauc=0.58755 val_adapt_rocauc=0.58315 adapt_steps=6.03 halt=0.16 train_steps=4.18
ep30 val_rocauc=0.65165 val_adapt_rocauc=0.65201 adapt_steps=7.28 halt=0.17 train_steps=4.07
ep40 val_rocauc=0.56667 val_adapt_rocauc=0.56703 adapt_steps=7.34 halt=0.19 train_steps=3.89
ep50 val_rocauc=0.56703 val_adapt_rocauc=0.55275 adapt_steps=5.61 halt=0.21 train_steps=3.62
ep60 val_rocauc=0.65348 val_adapt_rocauc=0.64286 adapt_steps=4.45 halt=0.28 train_steps=3.22
ep70 val_rocauc=0.60403 val_adapt_rocauc=0.60623 adapt_steps=6.13 halt=0.20 train_steps=3.81
ep80 val_rocauc=0.61685 val_adapt_rocauc=0.62051 adapt_steps=4.32 halt=0.29 train_steps=2.95
ep90 val_rocauc=0.62308 val_adapt_rocauc=0.63223 adapt_steps=4.03 halt=0.32 train_steps=2.75
ep100 val_rocauc=0.63150 val_adapt_rocauc=0.64066 adapt_steps=3.91 halt=0.31 train_steps=2.88
[ogbg-molbace_gen_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=60 val={'rocauc': 0.6534798534798535} test={'rocauc': 0.6915319074943488} adaptive={'rocauc': 0.6788384628760215} steps=3.9802631578947367
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbace_gen_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molbace view=film compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbace --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:0 --num_workers 0
ep10 val_rocauc=0.62637 val_adapt_rocauc=0.61355 adapt_steps=6.25 halt=0.18 train_steps=3.97
ep20 val_rocauc=0.57436 val_adapt_rocauc=0.57473 adapt_steps=7.62 halt=0.14 train_steps=4.26
ep30 val_rocauc=0.54359 val_adapt_rocauc=0.56227 adapt_steps=6.98 halt=0.16 train_steps=4.10
ep40 val_rocauc=0.59689 val_adapt_rocauc=0.58095 adapt_steps=6.34 halt=0.17 train_steps=4.15
ep50 val_rocauc=0.61868 val_adapt_rocauc=0.61648 adapt_steps=7.64 halt=0.16 train_steps=4.16
ep60 val_rocauc=0.59634 val_adapt_rocauc=0.59890 adapt_steps=6.07 halt=0.19 train_steps=3.90
ep70 val_rocauc=0.64670 val_adapt_rocauc=0.64725 adapt_steps=6.14 halt=0.22 train_steps=3.54
ep80 val_rocauc=0.59744 val_adapt_rocauc=0.59359 adapt_steps=5.54 halt=0.22 train_steps=3.55
ep90 val_rocauc=0.63333 val_adapt_rocauc=0.63187 adapt_steps=5.75 halt=0.23 train_steps=3.63
ep100 val_rocauc=0.62381 val_adapt_rocauc=0.61703 adapt_steps=5.77 halt=0.24 train_steps=3.54
[ogbg-molbace_film_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=70 val={'rocauc': 0.6467032967032967} test={'rocauc': 0.6802295252999478} adaptive={'rocauc': 0.6732742131803164} steps=5.256578947368421
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbace_film_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molbace view=resgated compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbace --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:0 --num_workers 0
ep10 val_rocauc=0.61245 val_adapt_rocauc=0.64286 adapt_steps=7.03 halt=0.17 train_steps=3.93
ep20 val_rocauc=0.53004 val_adapt_rocauc=0.55201 adapt_steps=5.85 halt=0.25 train_steps=3.12
ep30 val_rocauc=0.69158 val_adapt_rocauc=0.65897 adapt_steps=3.81 halt=0.27 train_steps=3.13
ep40 val_rocauc=0.58132 val_adapt_rocauc=0.57509 adapt_steps=4.23 halt=0.26 train_steps=3.03
ep50 val_rocauc=0.73956 val_adapt_rocauc=0.72930 adapt_steps=4.14 halt=0.26 train_steps=3.32
ep60 val_rocauc=0.67766 val_adapt_rocauc=0.67253 adapt_steps=5.34 halt=0.25 train_steps=3.32
ep70 val_rocauc=0.72418 val_adapt_rocauc=0.72967 adapt_steps=3.80 halt=0.30 train_steps=2.96
ep80 val_rocauc=0.71062 val_adapt_rocauc=0.69963 adapt_steps=3.62 halt=0.28 train_steps=3.28
ep90 val_rocauc=0.71538 val_adapt_rocauc=0.71612 adapt_steps=3.32 halt=0.34 train_steps=2.66
ep100 val_rocauc=0.69963 val_adapt_rocauc=0.69670 adapt_steps=3.08 halt=0.33 train_steps=2.65
[ogbg-molbace_resgated_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=50 val={'rocauc': 0.7395604395604396} test={'rocauc': 0.7313510693792384} adaptive={'rocauc': 0.7617805599026256} steps=3.388157894736842
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbace_resgated_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molbace view=tag compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbace --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:0 --num_workers 0
ep10 val_rocauc=0.65641 val_adapt_rocauc=0.66630 adapt_steps=5.21 halt=0.15 train_steps=4.34
ep20 val_rocauc=0.54725 val_adapt_rocauc=0.53040 adapt_steps=4.77 halt=0.20 train_steps=3.52
ep30 val_rocauc=0.51172 val_adapt_rocauc=0.58132 adapt_steps=4.25 halt=0.18 train_steps=3.92
ep40 val_rocauc=0.52821 val_adapt_rocauc=0.54835 adapt_steps=2.84 halt=0.30 train_steps=2.92
ep50 val_rocauc=0.62088 val_adapt_rocauc=0.63736 adapt_steps=3.09 halt=0.32 train_steps=2.79
ep60 val_rocauc=0.70842 val_adapt_rocauc=0.69707 adapt_steps=2.97 halt=0.36 train_steps=2.52
ep70 val_rocauc=0.68938 val_adapt_rocauc=0.72711 adapt_steps=3.27 halt=0.35 train_steps=2.61
ep80 val_rocauc=0.63516 val_adapt_rocauc=0.66667 adapt_steps=2.93 halt=0.39 train_steps=2.19
ep90 val_rocauc=0.62894 val_adapt_rocauc=0.65861 adapt_steps=2.70 halt=0.39 train_steps=2.17
ep100 val_rocauc=0.62930 val_adapt_rocauc=0.66154 adapt_steps=2.59 halt=0.40 train_steps=2.06
[ogbg-molbace_tag_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=60 val={'rocauc': 0.7084249084249085} test={'rocauc': 0.7960354720918101} adaptive={'rocauc': 0.7776038949747869} steps=2.5526315789473686
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbace_tag_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molbace view=sgc compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbace --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:0 --num_workers 0
ep10 val_rocauc=0.56667 val_adapt_rocauc=0.65934 adapt_steps=5.01 halt=0.16 train_steps=4.12
ep20 val_rocauc=0.62637 val_adapt_rocauc=0.63626 adapt_steps=5.57 halt=0.17 train_steps=4.08
ep30 val_rocauc=0.57985 val_adapt_rocauc=0.58498 adapt_steps=6.50 halt=0.21 train_steps=3.54
ep40 val_rocauc=0.64286 val_adapt_rocauc=0.64469 adapt_steps=6.27 halt=0.20 train_steps=3.71
ep50 val_rocauc=0.65678 val_adapt_rocauc=0.68791 adapt_steps=4.30 halt=0.25 train_steps=3.33
ep60 val_rocauc=0.66154 val_adapt_rocauc=0.66410 adapt_steps=3.97 halt=0.25 train_steps=3.40
ep70 val_rocauc=0.64799 val_adapt_rocauc=0.63223 adapt_steps=3.69 halt=0.27 train_steps=3.13
ep80 val_rocauc=0.63114 val_adapt_rocauc=0.62711 adapt_steps=4.02 halt=0.29 train_steps=3.02
ep90 val_rocauc=0.64103 val_adapt_rocauc=0.62125 adapt_steps=3.56 halt=0.31 train_steps=2.86
ep100 val_rocauc=0.63223 val_adapt_rocauc=0.61795 adapt_steps=3.49 halt=0.33 train_steps=2.77
[ogbg-molbace_sgc_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=60 val={'rocauc': 0.6615384615384615} test={'rocauc': 0.7970787689097547} adaptive={'rocauc': 0.7988175969396626} steps=3.638157894736842
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbace_sgc_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molbace view=cheb compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbace --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:0 --num_workers 0
ep10 val_rocauc=0.63846 val_adapt_rocauc=0.65495 adapt_steps=4.03 halt=0.18 train_steps=3.80
ep20 val_rocauc=0.67143 val_adapt_rocauc=0.69011 adapt_steps=4.54 halt=0.19 train_steps=3.98
ep30 val_rocauc=0.61245 val_adapt_rocauc=0.69084 adapt_steps=3.26 halt=0.28 train_steps=2.95
ep40 val_rocauc=0.59817 val_adapt_rocauc=0.60879 adapt_steps=3.16 halt=0.28 train_steps=3.20
ep50 val_rocauc=0.67729 val_adapt_rocauc=0.68498 adapt_steps=3.72 halt=0.33 train_steps=2.63
ep60 val_rocauc=0.65714 val_adapt_rocauc=0.68938 adapt_steps=3.75 halt=0.34 train_steps=2.55
ep70 val_rocauc=0.71648 val_adapt_rocauc=0.73810 adapt_steps=2.87 halt=0.41 train_steps=2.00
ep80 val_rocauc=0.70440 val_adapt_rocauc=0.70366 adapt_steps=2.56 halt=0.42 train_steps=2.02
ep90 val_rocauc=0.67656 val_adapt_rocauc=0.68315 adapt_steps=2.33 halt=0.46 train_steps=1.67
ep100 val_rocauc=0.67729 val_adapt_rocauc=0.68352 adapt_steps=2.35 halt=0.45 train_steps=1.70
[ogbg-molbace_cheb_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=70 val={'rocauc': 0.7164835164835166} test={'rocauc': 0.8007303077725613} adaptive={'rocauc': 0.8120326899669622} steps=2.598684210526316
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbace_cheb_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molbace view=arma compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbace --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:0 --num_workers 0
ep10 val_rocauc=0.62711 val_adapt_rocauc=0.60293 adapt_steps=4.62 halt=0.25 train_steps=3.40
ep20 val_rocauc=0.68095 val_adapt_rocauc=0.63114 adapt_steps=4.19 halt=0.27 train_steps=3.04
ep30 val_rocauc=0.53516 val_adapt_rocauc=0.52125 adapt_steps=6.76 halt=0.21 train_steps=3.47
ep40 val_rocauc=0.72564 val_adapt_rocauc=0.74139 adapt_steps=4.15 halt=0.26 train_steps=3.14
ep50 val_rocauc=0.73553 val_adapt_rocauc=0.73480 adapt_steps=4.72 halt=0.26 train_steps=3.28
ep60 val_rocauc=0.68278 val_adapt_rocauc=0.61465 adapt_steps=3.82 halt=0.30 train_steps=2.82
ep70 val_rocauc=0.70659 val_adapt_rocauc=0.71245 adapt_steps=3.50 halt=0.34 train_steps=2.66
ep80 val_rocauc=0.69853 val_adapt_rocauc=0.68718 adapt_steps=3.68 halt=0.34 train_steps=2.53
ep90 val_rocauc=0.69597 val_adapt_rocauc=0.68901 adapt_steps=2.83 halt=0.40 train_steps=2.18
ep100 val_rocauc=0.68828 val_adapt_rocauc=0.67766 adapt_steps=2.85 halt=0.39 train_steps=2.21
[ogbg-molbace_arma_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=50 val={'rocauc': 0.7355311355311356} test={'rocauc': 0.7450878108155102} adaptive={'rocauc': 0.7645626847504782} steps=4.375
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbace_arma_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molbace view=mf compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbace --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:0 --num_workers 0
ep10 val_rocauc=0.74066 val_adapt_rocauc=0.75311 adapt_steps=6.21 halt=0.21 train_steps=3.61
ep20 val_rocauc=0.68388 val_adapt_rocauc=0.68755 adapt_steps=6.04 halt=0.22 train_steps=3.45
ep30 val_rocauc=0.77473 val_adapt_rocauc=0.74725 adapt_steps=5.03 halt=0.19 train_steps=3.88
ep40 val_rocauc=0.63260 val_adapt_rocauc=0.63773 adapt_steps=5.52 halt=0.26 train_steps=3.11
ep50 val_rocauc=0.62894 val_adapt_rocauc=0.63260 adapt_steps=3.91 halt=0.29 train_steps=2.98
ep60 val_rocauc=0.70586 val_adapt_rocauc=0.69927 adapt_steps=3.87 halt=0.29 train_steps=2.99
ep70 val_rocauc=0.67656 val_adapt_rocauc=0.67619 adapt_steps=3.77 halt=0.33 train_steps=2.76
ep80 val_rocauc=0.68022 val_adapt_rocauc=0.67326 adapt_steps=3.08 halt=0.33 train_steps=2.72
ep90 val_rocauc=0.67179 val_adapt_rocauc=0.66740 adapt_steps=2.82 halt=0.38 train_steps=2.33
ep100 val_rocauc=0.67949 val_adapt_rocauc=0.68205 adapt_steps=2.88 halt=0.39 train_steps=2.19
[ogbg-molbace_mf_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=30 val={'rocauc': 0.7747252747252747} test={'rocauc': 0.8019474873934969} adaptive={'rocauc': 0.7949921752738653} steps=4.980263157894737
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbace_mf_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molbace view=appnp compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbace --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:0 --num_workers 0
ep10 val_rocauc=0.55678 val_adapt_rocauc=0.62051 adapt_steps=2.15 halt=0.39 train_steps=2.28
ep20 val_rocauc=0.68352 val_adapt_rocauc=0.65495 adapt_steps=2.36 halt=0.42 train_steps=1.96
ep30 val_rocauc=0.63700 val_adapt_rocauc=0.65092 adapt_steps=2.30 halt=0.44 train_steps=1.66
ep40 val_rocauc=0.61429 val_adapt_rocauc=0.60037 adapt_steps=2.01 halt=0.46 train_steps=1.65
ep50 val_rocauc=0.59414 val_adapt_rocauc=0.59011 adapt_steps=2.03 halt=0.47 train_steps=1.57
ep60 val_rocauc=0.61868 val_adapt_rocauc=0.59670 adapt_steps=2.00 halt=0.48 train_steps=1.59
ep70 val_rocauc=0.58901 val_adapt_rocauc=0.60256 adapt_steps=2.00 halt=0.47 train_steps=1.59
ep80 val_rocauc=0.56923 val_adapt_rocauc=0.58608 adapt_steps=2.00 halt=0.47 train_steps=1.62
ep90 val_rocauc=0.57839 val_adapt_rocauc=0.60183 adapt_steps=2.00 halt=0.48 train_steps=1.58
ep100 val_rocauc=0.58059 val_adapt_rocauc=0.60073 adapt_steps=2.00 halt=0.46 train_steps=1.60
[ogbg-molbace_appnp_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=20 val={'rocauc': 0.6835164835164835} test={'rocauc': 0.7118761954442705} adaptive={'rocauc': 0.7328290732046601} steps=2.1315789473684212
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbace_appnp_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
|
