-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathcomparison_log.txt
More file actions
336 lines (309 loc) · 42.9 KB
/
comparison_log.txt
File metadata and controls
336 lines (309 loc) · 42.9 KB
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
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
Note: mordred not installed. Using RDKit approximations for quantum descriptors.
Install with: pip install mordred
Using device: cpu
======================================================================
LOADING BBBP DATA (STANDARD - 15 features)
======================================================================
Loading BBBP dataset (quantum=False)...
Total samples: 2050
python.exe : [16:50:43] Explicit valence for atom # 1 N, 4, is greater than permitted
At line:1 char:43
+ ... BB_System'; & 'C:\Users\nakhi\anaconda3\python.exe' -u run_full_compa ...
+ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+ CategoryInfo : NotSpecified: ([16:50:43] Expl... than permitted:String) [], RemoteException
+ FullyQualifiedErrorId : NativeCommandError
[16:50:43] WARNING: not removing hydrogen atom without neighbors
[16:50:43] Explicit valence for atom # 6 N, 4, is greater than permitted
[16:50:43] WARNING: not removing hydrogen atom without neighbors
[16:50:43] WARNING: not removing hydrogen atom without neighbors
[16:50:43] WARNING: not removing hydrogen atom without neighbors
[16:50:43] WARNING: not removing hydrogen atom without neighbors
[16:50:43] WARNING: not removing hydrogen atom without neighbors
[16:50:43] WARNING: not removing hydrogen atom without neighbors
[16:50:43] Explicit valence for atom # 6 N, 4, is greater than permitted
[16:50:43] WARNING: not removing hydrogen atom without neighbors
[16:50:43] WARNING: not removing hydrogen atom without neighbors
[16:50:43] WARNING: not removing hydrogen atom without neighbors
[16:50:43] WARNING: not removing hydrogen atom without neighbors
[16:50:43] Explicit valence for atom # 11 N, 4, is greater than permitted
[16:50:43] Explicit valence for atom # 12 N, 4, is greater than permitted
[16:50:43] Explicit valence for atom # 5 N, 4, is greater than permitted
[16:50:43] Explicit valence for atom # 5 N, 4, is greater than permitted
[16:50:43] Explicit valence for atom # 5 N, 4, is greater than permitted
[16:50:43] Explicit valence for atom # 5 N, 4, is greater than permitted
[16:50:43] Explicit valence for atom # 5 N, 4, is greater than permitted
[16:50:43] WARNING: not removing hydrogen atom without neighbors
[16:50:43] WARNING: not removing hydrogen atom without neighbors
[16:50:43] Explicit valence for atom # 5 N, 4, is greater than permitted
[16:50:43] WARNING: not removing hydrogen atom without neighbors
[16:50:43] WARNING: not removing hydrogen atom without neighbors
[16:50:44] WARNING: not removing hydrogen atom without neighbors
[16:50:44] WARNING: not removing hydrogen atom without neighbors
[16:50:44] WARNING: not removing hydrogen atom without neighbors
[16:50:44] WARNING: not removing hydrogen atom without neighbors
[16:50:44] WARNING: not removing hydrogen atom without neighbors
[16:50:44] WARNING: not removing hydrogen atom without neighbors
[16:50:44] WARNING: not removing hydrogen atom without neighbors
[16:50:44] WARNING: not removing hydrogen atom without neighbors
[16:50:44] WARNING: not removing hydrogen atom without neighbors
[16:50:44] WARNING: not removing hydrogen atom without neighbors
[16:50:44] WARNING: not removing hydrogen atom without neighbors
[16:50:44] WARNING: not removing hydrogen atom without neighbors
[16:50:44] WARNING: not removing hydrogen atom without neighbors
[16:50:44] WARNING: not removing hydrogen atom without neighbors
[16:50:44] WARNING: not removing hydrogen atom without neighbors
[16:50:44] WARNING: not removing hydrogen atom without neighbors
[16:50:44] WARNING: not removing hydrogen atom without neighbors
[16:50:44] WARNING: not removing hydrogen atom without neighbors
[16:50:44] WARNING: not removing hydrogen atom without neighbors
[16:50:44] WARNING: not removing hydrogen atom without neighbors
[16:50:44] WARNING: not removing hydrogen atom without neighbors
[16:50:44] WARNING: not removing hydrogen atom without neighbors
[16:50:44] WARNING: not removing hydrogen atom without neighbors
[16:50:45] WARNING: not removing hydrogen atom without neighbors
[16:50:45] WARNING: not removing hydrogen atom without neighbors
Valid graphs: 2039
Features per node: 15
Train: 1631, Val: 204, Test: 204
======================================================================
MODEL 1: BASELINE (no pretraining, 15 features)
======================================================================
Training baseline_v2...
============================================================
Epoch 1/150 | Train AUC: 0.5130 | Val AUC: 0.7257 | Val Acc: 75.5% | LR: 0.000100 | *BEST*
Epoch 2/150 | Train AUC: 0.4915 | Val AUC: 0.7397 | Val Acc: 75.5% | LR: 0.000100 | *BEST*
Epoch 3/150 | Train AUC: 0.5554 | Val AUC: 0.7636 | Val Acc: 75.5% | LR: 0.000100 | *BEST*
Epoch 4/150 | Train AUC: 0.7248 | Val AUC: 0.8247 | Val Acc: 81.4% | LR: 0.000100 | *BEST*
Epoch 5/150 | Train AUC: 0.7556 | Val AUC: 0.8458 | Val Acc: 84.8% | LR: 0.000100 | *BEST*
Epoch 6/150 | Train AUC: 0.7621 | Val AUC: 0.8508 | Val Acc: 86.8% | LR: 0.000100 | *BEST*
Epoch 7/150 | Train AUC: 0.7503 | Val AUC: 0.8558 | Val Acc: 81.4% | LR: 0.000100 | *BEST*
Epoch 8/150 | Train AUC: 0.7920 | Val AUC: 0.8590 | Val Acc: 82.8% | LR: 0.000100 | *BEST*
Epoch 10/150 | Train AUC: 0.7821 | Val AUC: 0.8557 | Val Acc: 85.8% | LR: 0.000100
Epoch 11/150 | Train AUC: 0.7929 | Val AUC: 0.8604 | Val Acc: 85.3% | LR: 0.000100 | *BEST*
Epoch 12/150 | Train AUC: 0.8079 | Val AUC: 0.8629 | Val Acc: 85.8% | LR: 0.000100 | *BEST*
Epoch 18/150 | Train AUC: 0.8035 | Val AUC: 0.8629 | Val Acc: 85.3% | LR: 0.000100 | *BEST*
Epoch 20/150 | Train AUC: 0.8052 | Val AUC: 0.8601 | Val Acc: 84.3% | LR: 0.000100
Epoch 30/150 | Train AUC: 0.8168 | Val AUC: 0.8549 | Val Acc: 84.3% | LR: 0.000050
Epoch 40/150 | Train AUC: 0.8207 | Val AUC: 0.8371 | Val Acc: 79.9% | LR: 0.000025
Epoch 50/150 | Train AUC: 0.8346 | Val AUC: 0.8443 | Val Acc: 80.4% | LR: 0.000013
Early stopping at epoch 58. Best AUC: 0.8629 at epoch 18
Baseline Test Results: AUC=0.7677, Acc=79.9%
======================================================================
MODEL 2: PRETRAINED (ZINC 250k, 15 features)
======================================================================
Loading pretrained weights from models/pretrained_encoder.pth...
Transferred 5 layers, skipped 36 layers
Training pretrained_v2...
============================================================
Epoch 1/150 | Train AUC: 0.4774 | Val AUC: 0.8353 | Val Acc: 75.5% | LR: 0.000050 | *BEST*
Epoch 7/150 | Train AUC: 0.7802 | Val AUC: 0.8408 | Val Acc: 86.8% | LR: 0.000050 | *BEST*
Epoch 9/150 | Train AUC: 0.7951 | Val AUC: 0.8465 | Val Acc: 87.7% | LR: 0.000050 | *BEST*
Epoch 10/150 | Train AUC: 0.8037 | Val AUC: 0.8470 | Val Acc: 83.3% | LR: 0.000050 | *BEST*
Epoch 12/150 | Train AUC: 0.7810 | Val AUC: 0.8568 | Val Acc: 81.9% | LR: 0.000050 | *BEST*
Epoch 20/150 | Train AUC: 0.7991 | Val AUC: 0.8510 | Val Acc: 86.8% | LR: 0.000050
Epoch 23/150 | Train AUC: 0.8206 | Val AUC: 0.8592 | Val Acc: 84.8% | LR: 0.000050 | *BEST*
Epoch 24/150 | Train AUC: 0.8137 | Val AUC: 0.8665 | Val Acc: 85.3% | LR: 0.000050 | *BEST*
Epoch 30/150 | Train AUC: 0.8316 | Val AUC: 0.8536 | Val Acc: 82.8% | LR: 0.000050
Epoch 31/150 | Train AUC: 0.8307 | Val AUC: 0.8677 | Val Acc: 84.8% | LR: 0.000050 | *BEST*
Epoch 33/150 | Train AUC: 0.8422 | Val AUC: 0.8713 | Val Acc: 85.8% | LR: 0.000050 | *BEST*
Epoch 36/150 | Train AUC: 0.8364 | Val AUC: 0.8777 | Val Acc: 84.8% | LR: 0.000050 | *BEST*
Epoch 40/150 | Train AUC: 0.8573 | Val AUC: 0.8675 | Val Acc: 83.8% | LR: 0.000050
Epoch 43/150 | Train AUC: 0.8539 | Val AUC: 0.8886 | Val Acc: 84.3% | LR: 0.000050 | *BEST*
Epoch 50/150 | Train AUC: 0.8508 | Val AUC: 0.8657 | Val Acc: 83.8% | LR: 0.000050
Epoch 60/150 | Train AUC: 0.8776 | Val AUC: 0.8810 | Val Acc: 84.8% | LR: 0.000025
Epoch 70/150 | Train AUC: 0.8852 | Val AUC: 0.8768 | Val Acc: 85.3% | LR: 0.000013
Epoch 80/150 | Train AUC: 0.8962 | Val AUC: 0.8768 | Val Acc: 86.3% | LR: 0.000006
Early stopping at epoch 83. Best AUC: 0.8886 at epoch 43
Pretrained Test Results: AUC=0.7847, Acc=81.4%
======================================================================
LOADING BBBP DATA (QUANTUM - 28 features)
======================================================================
Loading BBBP dataset (quantum=True)...
Total samples: 2050
[19:37:30] Explicit valence for atom # 1 N, 4, is greater than permitted
[19:37:30] WARNING: not removing hydrogen atom without neighbors
[19:37:30] WARNING: not removing hydrogen atom without neighbors
[19:37:30] Explicit valence for atom # 6 N, 4, is greater than permitted
[19:37:31] WARNING: not removing hydrogen atom without neighbors
[19:37:31] WARNING: not removing hydrogen atom without neighbors
[19:37:31] WARNING: not removing hydrogen atom without neighbors
[19:37:31] WARNING: not removing hydrogen atom without neighbors
[19:37:31] WARNING: not removing hydrogen atom without neighbors
[19:37:31] WARNING: not removing hydrogen atom without neighbors
[19:37:31] WARNING: not removing hydrogen atom without neighbors
[19:37:31] WARNING: not removing hydrogen atom without neighbors
[19:37:31] WARNING: not removing hydrogen atom without neighbors
[19:37:31] WARNING: not removing hydrogen atom without neighbors
[19:37:31] WARNING: not removing hydrogen atom without neighbors
[19:37:31] WARNING: not removing hydrogen atom without neighbors
[19:37:32] Explicit valence for atom # 6 N, 4, is greater than permitted
[19:37:32] WARNING: not removing hydrogen atom without neighbors
[19:37:32] WARNING: not removing hydrogen atom without neighbors
[19:37:32] WARNING: not removing hydrogen atom without neighbors
[19:37:32] WARNING: not removing hydrogen atom without neighbors
[19:37:32] WARNING: not removing hydrogen atom without neighbors
[19:37:32] WARNING: not removing hydrogen atom without neighbors
[19:37:33] WARNING: not removing hydrogen atom without neighbors
[19:37:33] WARNING: not removing hydrogen atom without neighbors
[19:37:33] Explicit valence for atom # 11 N, 4, is greater than permitted
[19:37:33] Explicit valence for atom # 12 N, 4, is greater than permitted
[19:37:33] Explicit valence for atom # 5 N, 4, is greater than permitted
[19:37:33] Explicit valence for atom # 5 N, 4, is greater than permitted
[19:37:33] Explicit valence for atom # 5 N, 4, is greater than permitted
[19:37:33] Explicit valence for atom # 5 N, 4, is greater than permitted
[19:37:33] Explicit valence for atom # 5 N, 4, is greater than permitted
[19:37:33] WARNING: not removing hydrogen atom without neighbors
[19:37:33] WARNING: not removing hydrogen atom without neighbors
[19:37:33] WARNING: not removing hydrogen atom without neighbors
[19:37:33] WARNING: not removing hydrogen atom without neighbors
[19:37:33] Explicit valence for atom # 5 N, 4, is greater than permitted
[19:37:33] WARNING: not removing hydrogen atom without neighbors
[19:37:33] WARNING: not removing hydrogen atom without neighbors
[19:37:34] WARNING: not removing hydrogen atom without neighbors
[19:37:34] WARNING: not removing hydrogen atom without neighbors
[19:37:34] WARNING: not removing hydrogen atom without neighbors
[19:37:34] WARNING: not removing hydrogen atom without neighbors
[19:37:34] WARNING: not removing hydrogen atom without neighbors
[19:37:34] WARNING: not removing hydrogen atom without neighbors
[19:37:34] WARNING: not removing hydrogen atom without neighbors
[19:37:34] WARNING: not removing hydrogen atom without neighbors
[19:37:35] WARNING: not removing hydrogen atom without neighbors
[19:37:35] WARNING: not removing hydrogen atom without neighbors
[19:37:35] WARNING: not removing hydrogen atom without neighbors
[19:37:35] WARNING: not removing hydrogen atom without neighbors
[19:37:35] WARNING: not removing hydrogen atom without neighbors
[19:37:35] WARNING: not removing hydrogen atom without neighbors
[19:37:35] WARNING: not removing hydrogen atom without neighbors
[19:37:35] WARNING: not removing hydrogen atom without neighbors
[19:37:36] WARNING: not removing hydrogen atom without neighbors
[19:37:36] WARNING: not removing hydrogen atom without neighbors
[19:37:36] WARNING: not removing hydrogen atom without neighbors
[19:37:36] WARNING: not removing hydrogen atom without neighbors
[19:37:36] WARNING: not removing hydrogen atom without neighbors
[19:37:36] WARNING: not removing hydrogen atom without neighbors
[19:37:36] WARNING: not removing hydrogen atom without neighbors
[19:37:36] WARNING: not removing hydrogen atom without neighbors
[19:37:36] WARNING: not removing hydrogen atom without neighbors
[19:37:36] WARNING: not removing hydrogen atom without neighbors
[19:37:36] WARNING: not removing hydrogen atom without neighbors
[19:37:36] WARNING: not removing hydrogen atom without neighbors
[19:37:37] WARNING: not removing hydrogen atom without neighbors
[19:37:37] WARNING: not removing hydrogen atom without neighbors
[19:37:37] WARNING: not removing hydrogen atom without neighbors
[19:37:37] WARNING: not removing hydrogen atom without neighbors
[19:37:37] WARNING: not removing hydrogen atom without neighbors
[19:37:37] WARNING: not removing hydrogen atom without neighbors
[19:37:37] WARNING: not removing hydrogen atom without neighbors
[19:37:37] WARNING: not removing hydrogen atom without neighbors
[19:37:37] WARNING: not removing hydrogen atom without neighbors
[19:37:37] WARNING: not removing hydrogen atom without neighbors
[19:37:38] WARNING: not removing hydrogen atom without neighbors
[19:37:38] WARNING: not removing hydrogen atom without neighbors
[19:37:38] WARNING: not removing hydrogen atom without neighbors
[19:37:38] WARNING: not removing hydrogen atom without neighbors
[19:37:38] WARNING: not removing hydrogen atom without neighbors
[19:37:38] WARNING: not removing hydrogen atom without neighbors
[19:37:38] WARNING: not removing hydrogen atom without neighbors
[19:37:38] WARNING: not removing hydrogen atom without neighbors
[19:37:38] WARNING: not removing hydrogen atom without neighbors
[19:37:38] WARNING: not removing hydrogen atom without neighbors
[19:37:38] WARNING: not removing hydrogen atom without neighbors
[19:37:38] WARNING: not removing hydrogen atom without neighbors
[19:37:39] WARNING: not removing hydrogen atom without neighbors
[19:37:39] WARNING: not removing hydrogen atom without neighbors
Valid graphs: 2039
Features per node: 28
======================================================================
MODEL 3: QUANTUM-ONLY (28 features, no pretraining)
======================================================================
Training quantum_v2...
============================================================
Epoch 1/150 | Train AUC: 0.6970 | Val AUC: 0.8190 | Val Acc: 82.8% | LR: 0.000100 | *BEST*
Epoch 2/150 | Train AUC: 0.7706 | Val AUC: 0.8239 | Val Acc: 83.3% | LR: 0.000100 | *BEST*
Epoch 3/150 | Train AUC: 0.8002 | Val AUC: 0.8375 | Val Acc: 83.8% | LR: 0.000100 | *BEST*
Epoch 4/150 | Train AUC: 0.8034 | Val AUC: 0.8588 | Val Acc: 85.3% | LR: 0.000100 | *BEST*
Epoch 5/150 | Train AUC: 0.8111 | Val AUC: 0.8710 | Val Acc: 84.3% | LR: 0.000100 | *BEST*
Epoch 7/150 | Train AUC: 0.8110 | Val AUC: 0.8786 | Val Acc: 85.3% | LR: 0.000100 | *BEST*
Epoch 8/150 | Train AUC: 0.8256 | Val AUC: 0.8843 | Val Acc: 85.3% | LR: 0.000100 | *BEST*
Epoch 9/150 | Train AUC: 0.8165 | Val AUC: 0.8855 | Val Acc: 83.3% | LR: 0.000100 | *BEST*
Epoch 10/150 | Train AUC: 0.8287 | Val AUC: 0.8816 | Val Acc: 85.3% | LR: 0.000100
Epoch 11/150 | Train AUC: 0.8253 | Val AUC: 0.8864 | Val Acc: 84.8% | LR: 0.000100 | *BEST*
Epoch 12/150 | Train AUC: 0.8303 | Val AUC: 0.8917 | Val Acc: 85.3% | LR: 0.000100 | *BEST*
Epoch 13/150 | Train AUC: 0.8353 | Val AUC: 0.8951 | Val Acc: 86.3% | LR: 0.000100 | *BEST*
Epoch 14/150 | Train AUC: 0.8181 | Val AUC: 0.8978 | Val Acc: 84.8% | LR: 0.000100 | *BEST*
Epoch 15/150 | Train AUC: 0.8399 | Val AUC: 0.9066 | Val Acc: 85.3% | LR: 0.000100 | *BEST*
Epoch 16/150 | Train AUC: 0.8357 | Val AUC: 0.9094 | Val Acc: 86.3% | LR: 0.000100 | *BEST*
Epoch 18/150 | Train AUC: 0.8373 | Val AUC: 0.9161 | Val Acc: 86.8% | LR: 0.000100 | *BEST*
Epoch 20/150 | Train AUC: 0.8507 | Val AUC: 0.9129 | Val Acc: 86.3% | LR: 0.000100
Epoch 30/150 | Train AUC: 0.8584 | Val AUC: 0.9012 | Val Acc: 86.8% | LR: 0.000050
Epoch 40/150 | Train AUC: 0.8669 | Val AUC: 0.9031 | Val Acc: 86.3% | LR: 0.000025
Epoch 50/150 | Train AUC: 0.8720 | Val AUC: 0.9000 | Val Acc: 85.8% | LR: 0.000025
Early stopping at epoch 58. Best AUC: 0.9161 at epoch 18
Quantum Test Results: AUC=0.8240, Acc=78.9%
======================================================================
MODEL 4: PRETRAINED + QUANTUM (ZINC 250k pretraining + 28 features)
======================================================================
Loading pretrained weights from models/pretrained_encoder.pth...
Transferred 32 layers, skipped 9 layers
Training combined_v2...
============================================================
Epoch 1/150 | Train AUC: 0.6195 | Val AUC: 0.7335 | Val Acc: 79.4% | LR: 0.000050 | *BEST*
Epoch 2/150 | Train AUC: 0.7316 | Val AUC: 0.7849 | Val Acc: 80.9% | LR: 0.000050 | *BEST*
Epoch 3/150 | Train AUC: 0.7758 | Val AUC: 0.8147 | Val Acc: 81.9% | LR: 0.000050 | *BEST*
Epoch 4/150 | Train AUC: 0.7994 | Val AUC: 0.8387 | Val Acc: 83.3% | LR: 0.000050 | *BEST*
Epoch 5/150 | Train AUC: 0.8113 | Val AUC: 0.8561 | Val Acc: 84.3% | LR: 0.000050 | *BEST*
Epoch 6/150 | Train AUC: 0.8062 | Val AUC: 0.8626 | Val Acc: 86.8% | LR: 0.000050 | *BEST*
Epoch 8/150 | Train AUC: 0.8192 | Val AUC: 0.8636 | Val Acc: 85.8% | LR: 0.000050 | *BEST*
Epoch 9/150 | Train AUC: 0.8226 | Val AUC: 0.8670 | Val Acc: 85.8% | LR: 0.000050 | *BEST*
Epoch 10/150 | Train AUC: 0.8054 | Val AUC: 0.8830 | Val Acc: 85.3% | LR: 0.000050 | *BEST*
Epoch 17/150 | Train AUC: 0.8209 | Val AUC: 0.8849 | Val Acc: 86.8% | LR: 0.000050 | *BEST*
Epoch 20/150 | Train AUC: 0.8357 | Val AUC: 0.8705 | Val Acc: 87.7% | LR: 0.000050
Epoch 24/150 | Train AUC: 0.8186 | Val AUC: 0.8864 | Val Acc: 87.3% | LR: 0.000050 | *BEST*
Epoch 30/150 | Train AUC: 0.8244 | Val AUC: 0.8765 | Val Acc: 85.3% | LR: 0.000050
Epoch 31/150 | Train AUC: 0.8387 | Val AUC: 0.8869 | Val Acc: 86.3% | LR: 0.000050 | *BEST*
Epoch 34/150 | Train AUC: 0.8381 | Val AUC: 0.8878 | Val Acc: 85.3% | LR: 0.000050 | *BEST*
Epoch 35/150 | Train AUC: 0.8308 | Val AUC: 0.8892 | Val Acc: 85.3% | LR: 0.000050 | *BEST*
Epoch 36/150 | Train AUC: 0.8449 | Val AUC: 0.8896 | Val Acc: 87.3% | LR: 0.000050 | *BEST*
Epoch 40/150 | Train AUC: 0.8466 | Val AUC: 0.8921 | Val Acc: 85.3% | LR: 0.000050 | *BEST*
Epoch 45/150 | Train AUC: 0.8483 | Val AUC: 0.8922 | Val Acc: 85.8% | LR: 0.000050 | *BEST*
Epoch 46/150 | Train AUC: 0.8595 | Val AUC: 0.8942 | Val Acc: 85.3% | LR: 0.000050 | *BEST*
Epoch 48/150 | Train AUC: 0.8287 | Val AUC: 0.8951 | Val Acc: 86.8% | LR: 0.000050 | *BEST*
Epoch 49/150 | Train AUC: 0.8506 | Val AUC: 0.9031 | Val Acc: 85.8% | LR: 0.000050 | *BEST*
Epoch 50/150 | Train AUC: 0.8424 | Val AUC: 0.8994 | Val Acc: 85.8% | LR: 0.000050
Epoch 60/150 | Train AUC: 0.8574 | Val AUC: 0.8975 | Val Acc: 86.8% | LR: 0.000025
Epoch 63/150 | Train AUC: 0.8642 | Val AUC: 0.9035 | Val Acc: 85.3% | LR: 0.000025 | *BEST*
Epoch 66/150 | Train AUC: 0.8669 | Val AUC: 0.9039 | Val Acc: 85.3% | LR: 0.000025 | *BEST*
Epoch 70/150 | Train AUC: 0.8584 | Val AUC: 0.9017 | Val Acc: 85.8% | LR: 0.000025
Epoch 77/150 | Train AUC: 0.8585 | Val AUC: 0.9092 | Val Acc: 85.8% | LR: 0.000025 | *BEST*
Epoch 79/150 | Train AUC: 0.8711 | Val AUC: 0.9106 | Val Acc: 86.3% | LR: 0.000025 | *BEST*
Epoch 80/150 | Train AUC: 0.8793 | Val AUC: 0.9082 | Val Acc: 86.3% | LR: 0.000025
Epoch 89/150 | Train AUC: 0.8766 | Val AUC: 0.9119 | Val Acc: 85.8% | LR: 0.000025 | *BEST*
Epoch 90/150 | Train AUC: 0.8569 | Val AUC: 0.9071 | Val Acc: 84.8% | LR: 0.000025
Epoch 100/150 | Train AUC: 0.8798 | Val AUC: 0.9097 | Val Acc: 86.3% | LR: 0.000013
Epoch 103/150 | Train AUC: 0.8698 | Val AUC: 0.9123 | Val Acc: 84.8% | LR: 0.000013 | *BEST*
Epoch 110/150 | Train AUC: 0.8777 | Val AUC: 0.9100 | Val Acc: 84.3% | LR: 0.000013
Epoch 120/150 | Train AUC: 0.8704 | Val AUC: 0.9087 | Val Acc: 85.8% | LR: 0.000006
Epoch 130/150 | Train AUC: 0.8794 | Val AUC: 0.9051 | Val Acc: 84.8% | LR: 0.000003
Epoch 140/150 | Train AUC: 0.8832 | Val AUC: 0.9073 | Val Acc: 85.8% | LR: 0.000002
Early stopping at epoch 143. Best AUC: 0.9123 at epoch 103
Combined Test Results: AUC=0.8316, Acc=80.4%
======================================================================
FINAL 4-MODEL COMPARISON SUMMARY
======================================================================
Model Test AUC Accuracy Precision Recall F1
--------------------------------------------------------------------------------
Baseline 0.7677 79.9% 0.8654 0.8710 0.8682 (+0.0%)
Pretrained 0.7847 81.4% 0.8343 0.9419 0.8848 (+2.2%)
Quantum 0.8240 78.9% 0.8077 0.9484 0.8724 (+7.3%)
Pretrained Quantum 0.8316 80.4% 0.8249 0.9419 0.8795 (+8.3%)
--------------------------------------------------------------------------------
======================================================================
CATEGORY WINNERS
======================================================================
Best Overall (AUC): PRETRAINED_QUANTUM
AUC: 0.8316
Best Recall (finds most BBB+ compounds): QUANTUM
Recall: 0.9484
Best Precision (fewest false positives): BASELINE
Precision: 0.8654
Results saved to models/full_comparison_results_v2.npy