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-molbbbp view=gin compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbbbp --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=0.68724 val_adapt_rocauc=0.69058 adapt_steps=2.34 halt=0.32 train_steps=2.70
ep20 val_rocauc=0.71443 val_adapt_rocauc=0.73106 adapt_steps=4.76 halt=0.35 train_steps=2.43
ep30 val_rocauc=0.89704 val_adapt_rocauc=0.89107 adapt_steps=2.92 halt=0.32 train_steps=2.76
ep40 val_rocauc=0.92044 val_adapt_rocauc=0.93149 adapt_steps=3.13 halt=0.31 train_steps=2.82
ep50 val_rocauc=0.89326 val_adapt_rocauc=0.91984 adapt_steps=2.52 halt=0.34 train_steps=2.62
ep60 val_rocauc=0.88529 val_adapt_rocauc=0.90411 adapt_steps=2.35 halt=0.39 train_steps=2.28
ep70 val_rocauc=0.86876 val_adapt_rocauc=0.88728 adapt_steps=2.40 halt=0.38 train_steps=2.31
ep80 val_rocauc=0.90929 val_adapt_rocauc=0.91228 adapt_steps=2.47 halt=0.38 train_steps=2.27
ep90 val_rocauc=0.90491 val_adapt_rocauc=0.90760 adapt_steps=2.26 halt=0.42 train_steps=2.01
ep100 val_rocauc=0.90979 val_adapt_rocauc=0.91825 adapt_steps=2.25 halt=0.41 train_steps=2.05
[ogbg-molbbbp_gin_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=40 val={'rocauc': 0.92044209897441} test={'rocauc': 0.6527777777777778} adaptive={'rocauc': 0.65258487654321} steps=4.872549019607843
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbbbp_gin_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molbbbp view=gine compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbbbp --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=0.90073 val_adapt_rocauc=0.88450 adapt_steps=2.43 halt=0.27 train_steps=3.20
ep20 val_rocauc=0.92024 val_adapt_rocauc=0.93090 adapt_steps=3.02 halt=0.32 train_steps=2.63
ep30 val_rocauc=0.85622 val_adapt_rocauc=0.87673 adapt_steps=3.15 halt=0.34 train_steps=2.57
ep40 val_rocauc=0.89276 val_adapt_rocauc=0.88928 adapt_steps=3.26 halt=0.33 train_steps=2.70
ep50 val_rocauc=0.91795 val_adapt_rocauc=0.92114 adapt_steps=2.88 halt=0.34 train_steps=2.67
ep60 val_rocauc=0.91696 val_adapt_rocauc=0.90889 adapt_steps=2.46 halt=0.36 train_steps=2.44
ep70 val_rocauc=0.90501 val_adapt_rocauc=0.89037 adapt_steps=2.88 halt=0.38 train_steps=2.30
ep80 val_rocauc=0.92054 val_adapt_rocauc=0.91228 adapt_steps=2.57 halt=0.39 train_steps=2.24
ep90 val_rocauc=0.90949 val_adapt_rocauc=0.90312 adapt_steps=2.32 halt=0.41 train_steps=2.07
ep100 val_rocauc=0.91576 val_adapt_rocauc=0.91148 adapt_steps=2.29 halt=0.42 train_steps=2.02
[ogbg-molbbbp_gine_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=80 val={'rocauc': 0.9205416708154933} test={'rocauc': 0.591820987654321} adaptive={'rocauc': 0.5948109567901234} steps=2.9166666666666665
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbbbp_gine_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molbbbp view=gcn compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbbbp --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=0.82177 val_adapt_rocauc=0.93209 adapt_steps=2.11 halt=0.30 train_steps=2.97
ep20 val_rocauc=0.88241 val_adapt_rocauc=0.89923 adapt_steps=2.39 halt=0.38 train_steps=2.27
ep30 val_rocauc=0.84248 val_adapt_rocauc=0.89794 adapt_steps=2.54 halt=0.38 train_steps=2.19
ep40 val_rocauc=0.95280 val_adapt_rocauc=0.95390 adapt_steps=2.17 halt=0.41 train_steps=2.07
ep50 val_rocauc=0.89525 val_adapt_rocauc=0.92164 adapt_steps=2.07 halt=0.45 train_steps=1.74
ep60 val_rocauc=0.93727 val_adapt_rocauc=0.93050 adapt_steps=2.03 halt=0.45 train_steps=1.73
ep70 val_rocauc=0.92612 val_adapt_rocauc=0.92881 adapt_steps=2.03 halt=0.46 train_steps=1.68
ep80 val_rocauc=0.91845 val_adapt_rocauc=0.92273 adapt_steps=2.01 halt=0.47 train_steps=1.61
ep90 val_rocauc=0.91278 val_adapt_rocauc=0.91138 adapt_steps=2.00 halt=0.47 train_steps=1.66
ep100 val_rocauc=0.91477 val_adapt_rocauc=0.91626 adapt_steps=2.00 halt=0.47 train_steps=1.66
[ogbg-molbbbp_gcn_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=40 val={'rocauc': 0.9528029473264961} test={'rocauc': 0.6794945987654321} adaptive={'rocauc': 0.6629050925925926} steps=2.269607843137255
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbbbp_gcn_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molbbbp view=graphsage compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbbbp --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=0.94066 val_adapt_rocauc=0.95440 adapt_steps=2.54 halt=0.30 train_steps=2.86
ep20 val_rocauc=0.93299 val_adapt_rocauc=0.93956 adapt_steps=2.87 halt=0.38 train_steps=2.26
ep30 val_rocauc=0.83142 val_adapt_rocauc=0.88171 adapt_steps=2.41 halt=0.36 train_steps=2.40
ep40 val_rocauc=0.87494 val_adapt_rocauc=0.87374 adapt_steps=2.28 halt=0.40 train_steps=2.07
ep50 val_rocauc=0.92233 val_adapt_rocauc=0.94285 adapt_steps=2.27 halt=0.38 train_steps=2.38
ep60 val_rocauc=0.86478 val_adapt_rocauc=0.88539 adapt_steps=2.44 halt=0.40 train_steps=2.08
ep70 val_rocauc=0.91029 val_adapt_rocauc=0.91377 adapt_steps=2.14 halt=0.45 train_steps=1.75
ep80 val_rocauc=0.92532 val_adapt_rocauc=0.91218 adapt_steps=2.07 halt=0.46 train_steps=1.68
ep90 val_rocauc=0.90630 val_adapt_rocauc=0.90202 adapt_steps=2.06 halt=0.45 train_steps=1.73
ep100 val_rocauc=0.90531 val_adapt_rocauc=0.90083 adapt_steps=2.04 halt=0.46 train_steps=1.67
[ogbg-molbbbp_graphsage_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=10 val={'rocauc': 0.9406551827143284} test={'rocauc': 0.6039737654320988} adaptive={'rocauc': 0.6147762345679012} steps=2.5686274509803924
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbbbp_graphsage_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molbbbp view=gatv2 compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbbbp --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=0.87544 val_adapt_rocauc=0.89704 adapt_steps=3.72 halt=0.33 train_steps=2.64
ep20 val_rocauc=0.92492 val_adapt_rocauc=0.94852 adapt_steps=2.61 halt=0.39 train_steps=2.17
ep30 val_rocauc=0.89804 val_adapt_rocauc=0.93239 adapt_steps=2.27 halt=0.37 train_steps=2.35
ep40 val_rocauc=0.90511 val_adapt_rocauc=0.92612 adapt_steps=2.27 halt=0.41 train_steps=2.00
ep50 val_rocauc=0.90103 val_adapt_rocauc=0.89475 adapt_steps=2.46 halt=0.41 train_steps=2.01
ep60 val_rocauc=0.90401 val_adapt_rocauc=0.89306 adapt_steps=2.24 halt=0.41 train_steps=2.05
ep70 val_rocauc=0.93030 val_adapt_rocauc=0.91656 adapt_steps=2.10 halt=0.46 train_steps=1.67
ep80 val_rocauc=0.92233 val_adapt_rocauc=0.90411 adapt_steps=2.09 halt=0.46 train_steps=1.66
ep90 val_rocauc=0.91546 val_adapt_rocauc=0.89386 adapt_steps=2.11 halt=0.46 train_steps=1.70
ep100 val_rocauc=0.91287 val_adapt_rocauc=0.88868 adapt_steps=2.09 halt=0.47 train_steps=1.65
[ogbg-molbbbp_gatv2_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=70 val={'rocauc': 0.9302997112416609} test={'rocauc': 0.6648341049382716} adaptive={'rocauc': 0.6627121913580247} steps=2.1372549019607843
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbbbp_gatv2_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molbbbp view=graphconv compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbbbp --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=0.83053 val_adapt_rocauc=0.84875 adapt_steps=3.46 halt=0.34 train_steps=2.62
ep20 val_rocauc=0.71931 val_adapt_rocauc=0.75465 adapt_steps=3.00 halt=0.37 train_steps=2.40
ep30 val_rocauc=0.88380 val_adapt_rocauc=0.89704 adapt_steps=2.44 halt=0.37 train_steps=2.40
ep40 val_rocauc=0.89913 val_adapt_rocauc=0.91507 adapt_steps=2.17 halt=0.37 train_steps=2.35
ep50 val_rocauc=0.88938 val_adapt_rocauc=0.93737 adapt_steps=2.25 halt=0.43 train_steps=1.90
ep60 val_rocauc=0.92343 val_adapt_rocauc=0.93757 adapt_steps=2.21 halt=0.41 train_steps=2.04
ep70 val_rocauc=0.87703 val_adapt_rocauc=0.90889 adapt_steps=2.45 halt=0.41 train_steps=1.97
ep80 val_rocauc=0.89396 val_adapt_rocauc=0.91726 adapt_steps=2.09 halt=0.46 train_steps=1.73
ep90 val_rocauc=0.86498 val_adapt_rocauc=0.89167 adapt_steps=2.19 halt=0.46 train_steps=1.70
ep100 val_rocauc=0.87016 val_adapt_rocauc=0.89545 adapt_steps=2.13 halt=0.47 train_steps=1.68
[ogbg-molbbbp_graphconv_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=60 val={'rocauc': 0.9234292542069104} test={'rocauc': 0.6602044753086419} adaptive={'rocauc': 0.6875} steps=2.7941176470588234
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbbbp_graphconv_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molbbbp view=transformer compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbbbp --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=0.87494 val_adapt_rocauc=0.89316 adapt_steps=3.50 halt=0.33 train_steps=2.53
ep20 val_rocauc=0.88878 val_adapt_rocauc=0.89784 adapt_steps=2.60 halt=0.35 train_steps=2.46
ep30 val_rocauc=0.85243 val_adapt_rocauc=0.93747 adapt_steps=2.47 halt=0.36 train_steps=2.40
ep40 val_rocauc=0.93687 val_adapt_rocauc=0.94046 adapt_steps=2.35 halt=0.41 train_steps=2.09
ep50 val_rocauc=0.90959 val_adapt_rocauc=0.91875 adapt_steps=2.32 halt=0.41 train_steps=2.16
ep60 val_rocauc=0.93090 val_adapt_rocauc=0.93617 adapt_steps=2.40 halt=0.42 train_steps=1.90
ep70 val_rocauc=0.93199 val_adapt_rocauc=0.92134 adapt_steps=2.23 halt=0.45 train_steps=1.72
ep80 val_rocauc=0.93637 val_adapt_rocauc=0.93866 adapt_steps=2.15 halt=0.45 train_steps=1.78
ep90 val_rocauc=0.93976 val_adapt_rocauc=0.93378 adapt_steps=2.13 halt=0.45 train_steps=1.81
ep100 val_rocauc=0.93966 val_adapt_rocauc=0.93369 adapt_steps=2.13 halt=0.45 train_steps=1.79
[ogbg-molbbbp_transformer_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=90 val={'rocauc': 0.9397590361445782} test={'rocauc': 0.642554012345679} adaptive={'rocauc': 0.6309799382716049} steps=2.2254901960784315
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbbbp_transformer_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molbbbp view=pna compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbbbp --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=0.92731 val_adapt_rocauc=0.92950 adapt_steps=6.11 halt=0.24 train_steps=3.52
/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=0.89097 val_adapt_rocauc=0.90093 adapt_steps=2.08 halt=0.32 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(
ep30 val_rocauc=0.92433 val_adapt_rocauc=0.92502 adapt_steps=3.31 halt=0.34 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(
ep40 val_rocauc=0.93777 val_adapt_rocauc=0.94056 adapt_steps=2.97 halt=0.35 train_steps=2.49
/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=0.95001 val_adapt_rocauc=0.95051 adapt_steps=3.67 halt=0.31 train_steps=2.90
/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=0.91317 val_adapt_rocauc=0.89913 adapt_steps=2.84 halt=0.37 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=0.94583 val_adapt_rocauc=0.94613 adapt_steps=2.63 halt=0.36 train_steps=2.49
/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=0.90770 val_adapt_rocauc=0.90132 adapt_steps=2.44 halt=0.39 train_steps=2.22
/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=0.92761 val_adapt_rocauc=0.92482 adapt_steps=2.29 halt=0.42 train_steps=1.98
/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(
ep100 val_rocauc=0.92871 val_adapt_rocauc=0.92612 adapt_steps=2.34 halt=0.43 train_steps=1.94
[ogbg-molbbbp_pna_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=50 val={'rocauc': 0.9500149357761626} test={'rocauc': 0.6172839506172839} adaptive={'rocauc': 0.6179591049382716} steps=4.686274509803922
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbbbp_pna_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molbbbp view=gen compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbbbp --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=0.78104 val_adapt_rocauc=0.86747 adapt_steps=2.68 halt=0.35 train_steps=2.43
ep20 val_rocauc=0.89664 val_adapt_rocauc=0.89754 adapt_steps=4.28 halt=0.31 train_steps=2.84
ep30 val_rocauc=0.92512 val_adapt_rocauc=0.92114 adapt_steps=4.62 halt=0.30 train_steps=3.15
ep40 val_rocauc=0.88798 val_adapt_rocauc=0.86727 adapt_steps=3.66 halt=0.35 train_steps=2.40
ep50 val_rocauc=0.92383 val_adapt_rocauc=0.93737 adapt_steps=2.48 halt=0.35 train_steps=2.51
ep60 val_rocauc=0.91218 val_adapt_rocauc=0.91278 adapt_steps=2.67 halt=0.37 train_steps=2.35
ep70 val_rocauc=0.90083 val_adapt_rocauc=0.91785 adapt_steps=2.15 halt=0.40 train_steps=2.04
ep80 val_rocauc=0.91646 val_adapt_rocauc=0.92612 adapt_steps=2.24 halt=0.43 train_steps=1.95
ep90 val_rocauc=0.91616 val_adapt_rocauc=0.92751 adapt_steps=2.23 halt=0.45 train_steps=1.79
ep100 val_rocauc=0.90909 val_adapt_rocauc=0.91865 adapt_steps=2.24 halt=0.44 train_steps=1.86
[ogbg-molbbbp_gen_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=30 val={'rocauc': 0.925121975505327} test={'rocauc': 0.6313657407407408} adaptive={'rocauc': 0.6291473765432098} steps=3.7892156862745097
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbbbp_gen_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molbbbp view=film compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbbbp --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=0.65558 val_adapt_rocauc=0.57483 adapt_steps=4.25 halt=0.29 train_steps=2.83
ep20 val_rocauc=0.89615 val_adapt_rocauc=0.88997 adapt_steps=4.61 halt=0.29 train_steps=3.15
ep30 val_rocauc=0.91606 val_adapt_rocauc=0.91606 adapt_steps=3.94 halt=0.30 train_steps=2.97
ep40 val_rocauc=0.87683 val_adapt_rocauc=0.88221 adapt_steps=3.41 halt=0.27 train_steps=3.07
ep50 val_rocauc=0.93926 val_adapt_rocauc=0.94832 adapt_steps=2.70 halt=0.29 train_steps=3.03
ep60 val_rocauc=0.92204 val_adapt_rocauc=0.92881 adapt_steps=3.16 halt=0.29 train_steps=2.99
ep70 val_rocauc=0.89714 val_adapt_rocauc=0.90451 adapt_steps=2.59 halt=0.34 train_steps=2.64
ep80 val_rocauc=0.91835 val_adapt_rocauc=0.92413 adapt_steps=2.57 halt=0.35 train_steps=2.65
ep90 val_rocauc=0.91048 val_adapt_rocauc=0.91745 adapt_steps=2.50 halt=0.34 train_steps=2.57
ep100 val_rocauc=0.92582 val_adapt_rocauc=0.93189 adapt_steps=2.41 halt=0.37 train_steps=2.51
[ogbg-molbbbp_film_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=50 val={'rocauc': 0.9392611769391616} test={'rocauc': 0.5584490740740741} adaptive={'rocauc': 0.5745563271604939} steps=3.2058823529411766
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbbbp_film_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molbbbp view=resgated compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbbbp --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=0.79478 val_adapt_rocauc=0.79906 adapt_steps=4.00 halt=0.30 train_steps=2.83
ep20 val_rocauc=0.90192 val_adapt_rocauc=0.90431 adapt_steps=2.94 halt=0.36 train_steps=2.35
ep30 val_rocauc=0.91327 val_adapt_rocauc=0.93100 adapt_steps=2.63 halt=0.36 train_steps=2.32
ep40 val_rocauc=0.88051 val_adapt_rocauc=0.89535 adapt_steps=2.30 halt=0.37 train_steps=2.38
ep50 val_rocauc=0.90401 val_adapt_rocauc=0.90829 adapt_steps=2.96 halt=0.37 train_steps=2.33
ep60 val_rocauc=0.86060 val_adapt_rocauc=0.87822 adapt_steps=2.16 halt=0.41 train_steps=2.03
ep70 val_rocauc=0.90859 val_adapt_rocauc=0.90770 adapt_steps=2.04 halt=0.44 train_steps=1.88
ep80 val_rocauc=0.89893 val_adapt_rocauc=0.90581 adapt_steps=2.04 halt=0.45 train_steps=1.76
ep90 val_rocauc=0.89824 val_adapt_rocauc=0.89893 adapt_steps=2.02 halt=0.45 train_steps=1.72
ep100 val_rocauc=0.89206 val_adapt_rocauc=0.89406 adapt_steps=2.01 halt=0.45 train_steps=1.74
[ogbg-molbbbp_resgated_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=30 val={'rocauc': 0.9132729264164094} test={'rocauc': 0.5753279320987655} adaptive={'rocauc': 0.5887345679012346} steps=3.593137254901961
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbbbp_resgated_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molbbbp view=tag compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbbbp --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=0.79010 val_adapt_rocauc=0.90242 adapt_steps=2.69 halt=0.30 train_steps=2.96
ep20 val_rocauc=0.78353 val_adapt_rocauc=0.88868 adapt_steps=2.15 halt=0.32 train_steps=2.81
ep30 val_rocauc=0.94344 val_adapt_rocauc=0.94145 adapt_steps=2.73 halt=0.38 train_steps=2.31
ep40 val_rocauc=0.86787 val_adapt_rocauc=0.89565 adapt_steps=2.29 halt=0.39 train_steps=2.18
ep50 val_rocauc=0.94962 val_adapt_rocauc=0.95420 adapt_steps=2.25 halt=0.40 train_steps=2.20
ep60 val_rocauc=0.93269 val_adapt_rocauc=0.94374 adapt_steps=2.16 halt=0.41 train_steps=2.00
ep70 val_rocauc=0.93996 val_adapt_rocauc=0.94155 adapt_steps=2.23 halt=0.42 train_steps=1.93
ep80 val_rocauc=0.93727 val_adapt_rocauc=0.93598 adapt_steps=2.08 halt=0.46 train_steps=1.75
ep90 val_rocauc=0.93508 val_adapt_rocauc=0.92691 adapt_steps=2.04 halt=0.46 train_steps=1.73
ep100 val_rocauc=0.93717 val_adapt_rocauc=0.92980 adapt_steps=2.02 halt=0.46 train_steps=1.69
[ogbg-molbbbp_tag_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=50 val={'rocauc': 0.9496166484118291} test={'rocauc': 0.6579861111111112} adaptive={'rocauc': 0.6550925925925926} steps=2.6176470588235294
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbbbp_tag_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molbbbp view=sgc compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbbbp --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=0.65259 val_adapt_rocauc=0.79578 adapt_steps=2.69 halt=0.32 train_steps=2.68
ep20 val_rocauc=0.92612 val_adapt_rocauc=0.94066 adapt_steps=2.18 halt=0.38 train_steps=2.24
ep30 val_rocauc=0.93478 val_adapt_rocauc=0.94484 adapt_steps=2.27 halt=0.34 train_steps=2.50
ep40 val_rocauc=0.91576 val_adapt_rocauc=0.94056 adapt_steps=2.19 halt=0.42 train_steps=2.00
ep50 val_rocauc=0.90760 val_adapt_rocauc=0.92552 adapt_steps=2.36 halt=0.41 train_steps=2.02
ep60 val_rocauc=0.94753 val_adapt_rocauc=0.95798 adapt_steps=2.07 halt=0.45 train_steps=1.74
ep70 val_rocauc=0.93647 val_adapt_rocauc=0.94066 adapt_steps=2.13 halt=0.46 train_steps=1.65
ep80 val_rocauc=0.93289 val_adapt_rocauc=0.93777 adapt_steps=2.06 halt=0.46 train_steps=1.67
ep90 val_rocauc=0.93817 val_adapt_rocauc=0.94165 adapt_steps=2.05 halt=0.47 train_steps=1.68
ep100 val_rocauc=0.93757 val_adapt_rocauc=0.93916 adapt_steps=2.03 halt=0.46 train_steps=1.68
[ogbg-molbbbp_sgc_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=60 val={'rocauc': 0.947525639749079} test={'rocauc': 0.6593364197530864} adaptive={'rocauc': 0.6675347222222222} steps=2.2549019607843137
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbbbp_sgc_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molbbbp view=cheb compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbbbp --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=0.79966 val_adapt_rocauc=0.83352 adapt_steps=2.95 halt=0.33 train_steps=2.69
ep20 val_rocauc=0.87763 val_adapt_rocauc=0.88738 adapt_steps=2.30 halt=0.38 train_steps=2.38
ep30 val_rocauc=0.83511 val_adapt_rocauc=0.87225 adapt_steps=2.24 halt=0.40 train_steps=2.04
ep40 val_rocauc=0.93578 val_adapt_rocauc=0.94066 adapt_steps=2.38 halt=0.41 train_steps=2.09
ep50 val_rocauc=0.92881 val_adapt_rocauc=0.92642 adapt_steps=2.17 halt=0.44 train_steps=1.85
ep60 val_rocauc=0.92353 val_adapt_rocauc=0.91337 adapt_steps=2.12 halt=0.45 train_steps=1.78
ep70 val_rocauc=0.94962 val_adapt_rocauc=0.94494 adapt_steps=2.03 halt=0.46 train_steps=1.65
ep80 val_rocauc=0.93986 val_adapt_rocauc=0.93508 adapt_steps=2.13 halt=0.47 train_steps=1.65
ep90 val_rocauc=0.93080 val_adapt_rocauc=0.91835 adapt_steps=2.03 halt=0.46 train_steps=1.72
ep100 val_rocauc=0.92433 val_adapt_rocauc=0.91208 adapt_steps=2.03 halt=0.45 train_steps=1.80
[ogbg-molbbbp_cheb_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=70 val={'rocauc': 0.9496166484118291} test={'rocauc': 0.6182484567901234} adaptive={'rocauc': 0.611014660493827} steps=2.1666666666666665
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbbbp_cheb_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molbbbp view=arma compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbbbp --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=0.55989 val_adapt_rocauc=0.61585 adapt_steps=2.59 halt=0.37 train_steps=2.15
ep20 val_rocauc=0.88748 val_adapt_rocauc=0.93000 adapt_steps=2.67 halt=0.40 train_steps=2.09
ep30 val_rocauc=0.90172 val_adapt_rocauc=0.91427 adapt_steps=2.22 halt=0.39 train_steps=2.27
ep40 val_rocauc=0.93269 val_adapt_rocauc=0.93189 adapt_steps=2.33 halt=0.38 train_steps=2.31
ep50 val_rocauc=0.92771 val_adapt_rocauc=0.92433 adapt_steps=2.20 halt=0.46 train_steps=1.75
ep60 val_rocauc=0.91228 val_adapt_rocauc=0.91984 adapt_steps=2.14 halt=0.44 train_steps=1.83
ep70 val_rocauc=0.94046 val_adapt_rocauc=0.93757 adapt_steps=2.02 halt=0.45 train_steps=1.70
ep80 val_rocauc=0.92273 val_adapt_rocauc=0.92801 adapt_steps=2.06 halt=0.47 train_steps=1.64
ep90 val_rocauc=0.92154 val_adapt_rocauc=0.92542 adapt_steps=2.08 halt=0.45 train_steps=1.79
ep100 val_rocauc=0.91845 val_adapt_rocauc=0.92233 adapt_steps=2.07 halt=0.46 train_steps=1.74
[ogbg-molbbbp_arma_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=70 val={'rocauc': 0.9404560390321618} test={'rocauc': 0.6411072530864197} adaptive={'rocauc': 0.636766975308642} steps=2.1519607843137254
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbbbp_arma_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molbbbp view=mf compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbbbp --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=0.70109 val_adapt_rocauc=0.81280 adapt_steps=2.37 halt=0.29 train_steps=2.98
ep20 val_rocauc=0.91815 val_adapt_rocauc=0.92184 adapt_steps=3.19 halt=0.33 train_steps=2.59
ep30 val_rocauc=0.93030 val_adapt_rocauc=0.94095 adapt_steps=2.34 halt=0.36 train_steps=2.50
ep40 val_rocauc=0.93677 val_adapt_rocauc=0.94534 adapt_steps=2.69 halt=0.35 train_steps=2.51
ep50 val_rocauc=0.91098 val_adapt_rocauc=0.91238 adapt_steps=2.40 halt=0.38 train_steps=2.23
ep60 val_rocauc=0.92184 val_adapt_rocauc=0.92672 adapt_steps=2.20 halt=0.44 train_steps=1.87
ep70 val_rocauc=0.93598 val_adapt_rocauc=0.93926 adapt_steps=2.13 halt=0.45 train_steps=1.70
ep80 val_rocauc=0.90790 val_adapt_rocauc=0.91785 adapt_steps=2.10 halt=0.47 train_steps=1.64
ep90 val_rocauc=0.89864 val_adapt_rocauc=0.91019 adapt_steps=2.09 halt=0.47 train_steps=1.65
ep100 val_rocauc=0.89406 val_adapt_rocauc=0.90561 adapt_steps=2.11 halt=0.46 train_steps=1.70
[ogbg-molbbbp_mf_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=40 val={'rocauc': 0.936771880912078} test={'rocauc': 0.662133487654321} adaptive={'rocauc': 0.667824074074074} steps=3.5441176470588234
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbbbp_mf_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
[run] ogbg-molbbbp view=appnp compute=rrog-act mode=stream T=1 ns=3 seed=0 device=cuda:0
python3 rrog/train_ogb_graphprop.py --dataset ogbg-molbbbp --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=0.84975 val_adapt_rocauc=0.88719 adapt_steps=2.18 halt=0.43 train_steps=1.80
ep20 val_rocauc=0.88748 val_adapt_rocauc=0.92572 adapt_steps=2.08 halt=0.47 train_steps=1.62
ep30 val_rocauc=0.78114 val_adapt_rocauc=0.87962 adapt_steps=2.04 halt=0.46 train_steps=1.63
ep40 val_rocauc=0.86936 val_adapt_rocauc=0.91825 adapt_steps=2.01 halt=0.47 train_steps=1.63
ep50 val_rocauc=0.90660 val_adapt_rocauc=0.93378 adapt_steps=2.00 halt=0.46 train_steps=1.65
ep60 val_rocauc=0.82316 val_adapt_rocauc=0.90620 adapt_steps=2.01 halt=0.46 train_steps=1.68
ep70 val_rocauc=0.86389 val_adapt_rocauc=0.90242 adapt_steps=2.00 halt=0.46 train_steps=1.70
ep80 val_rocauc=0.82684 val_adapt_rocauc=0.89884 adapt_steps=2.02 halt=0.46 train_steps=1.68
ep90 val_rocauc=0.81639 val_adapt_rocauc=0.89884 adapt_steps=2.01 halt=0.47 train_steps=1.70
ep100 val_rocauc=0.82445 val_adapt_rocauc=0.90192 adapt_steps=2.01 halt=0.46 train_steps=1.64
[ogbg-molbbbp_appnp_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0] best_ep=50 val={'rocauc': 0.9066016130638256} test={'rocauc': 0.5573881172839507} adaptive={'rocauc': 0.6225887345679013} steps=2.014705882352941
wrote /orion/u/oscarwan/rrog-gnn-runner/runs/ogbg-molbbbp_appnp_rrog-act_T1_ns3_stream_hm8_hmin2_loss0.2_lq0.1_hex0.1_qw0_h128_e100_s0.json
|
