AbstractPhil commited on
Commit
4d13d43
·
verified ·
1 Parent(s): 3585cb2

Ablation I-LOW-I1_1layer

Browse files
I/LOW/I1_1layer/seed0/final_report.json ADDED
@@ -0,0 +1,204 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "config": {
3
+ "group": "I",
4
+ "variant": "I1_1layer",
5
+ "band": "LOW",
6
+ "seed": 0,
7
+ "phase": 1,
8
+ "overrides": {
9
+ "n_cross": 1,
10
+ "max_alpha": 0.2
11
+ },
12
+ "description": "I-LOW-I1_1layer"
13
+ },
14
+ "run_config": {
15
+ "matrix_v": 64,
16
+ "D": 16,
17
+ "patch_size": 16,
18
+ "hidden": 64,
19
+ "depth": 1,
20
+ "n_cross_layers": 1,
21
+ "n_heads": 4,
22
+ "max_alpha": 0.2,
23
+ "alpha_init": -2.0,
24
+ "img_size": 64,
25
+ "batch_size": 128,
26
+ "lr": 0.0001,
27
+ "epochs": 1,
28
+ "weight_decay": 0.0,
29
+ "use_cv_ema": true,
30
+ "cv_ema_alpha": 0.01,
31
+ "cv_alignment_epochs": 0,
32
+ "cv_measure_every": 50,
33
+ "cv_sigma_scale": 0.3,
34
+ "boost": 0.5,
35
+ "cross_attn_clip": 0.5,
36
+ "allowed_types": [
37
+ 0,
38
+ 1,
39
+ 2,
40
+ 3,
41
+ 4,
42
+ 5,
43
+ 6,
44
+ 7,
45
+ 8,
46
+ 9,
47
+ 10,
48
+ 11,
49
+ 12,
50
+ 13,
51
+ 14,
52
+ 15
53
+ ],
54
+ "train_size": 1000000,
55
+ "val_size": 10000,
56
+ "num_workers": 2,
57
+ "report_every": 100,
58
+ "major_report_every": 10,
59
+ "save_every": 5,
60
+ "seed": 0,
61
+ "hf_repo": "AbstractPhil/geolip-svae-batteries",
62
+ "upload": false
63
+ },
64
+ "cv_ema_final": 0.20322039627440905,
65
+ "cv_last": 0.20856704538854393,
66
+ "predicted_band": "LOW",
67
+ "expected_band": "LOW",
68
+ "band_match": true,
69
+ "test_mse": 1.0105767250061035,
70
+ "recon_ema": 1.0138646476168225,
71
+ "S0": 2.854369878768921,
72
+ "SD": 1.1607599258422852,
73
+ "ratio": 2.459052721093184,
74
+ "erank": 15.512493133544922,
75
+ "observed_sphere_cv": 0.19777425503589818,
76
+ "uniform_sphere_cv_prediction": 0.19902003258265127,
77
+ "band_deviation": -0.0012457775467530907,
78
+ "params_count": 250211,
79
+ "wallclock_seconds": 14.238457918167114,
80
+ "batches_completed": 1000,
81
+ "batch_limit": 1000,
82
+ "cv_trajectory": [
83
+ {
84
+ "batch": 0,
85
+ "cv": 0.20395987159654982,
86
+ "cv_ema": 0.20395987159654982,
87
+ "recon": 1.6308250427246094
88
+ },
89
+ {
90
+ "batch": 50,
91
+ "cv": 0.19974706666730166,
92
+ "cv_ema": 0.20391774354725734,
93
+ "recon": 1.6488956212997437
94
+ },
95
+ {
96
+ "batch": 100,
97
+ "cv": 0.205418092270394,
98
+ "cv_ema": 0.2039327470344887,
99
+ "recon": 1.3726505041122437
100
+ },
101
+ {
102
+ "batch": 150,
103
+ "cv": 0.1959209106218471,
104
+ "cv_ema": 0.20385262867036227,
105
+ "recon": 1.369863510131836
106
+ },
107
+ {
108
+ "batch": 200,
109
+ "cv": 0.20824419636311334,
110
+ "cv_ema": 0.2038965443472898,
111
+ "recon": 1.2076982259750366
112
+ },
113
+ {
114
+ "batch": 250,
115
+ "cv": 0.20550173413329773,
116
+ "cv_ema": 0.20391259624514987,
117
+ "recon": 1.1482267379760742
118
+ },
119
+ {
120
+ "batch": 300,
121
+ "cv": 0.1803511675647411,
122
+ "cv_ema": 0.20367698195834577,
123
+ "recon": 0.9758987426757812
124
+ },
125
+ {
126
+ "batch": 350,
127
+ "cv": 0.20185841382424027,
128
+ "cv_ema": 0.20365879627700473,
129
+ "recon": 0.9563150405883789
130
+ },
131
+ {
132
+ "batch": 400,
133
+ "cv": 0.1926903916743087,
134
+ "cv_ema": 0.20354911223097777,
135
+ "recon": 0.9693193435668945
136
+ },
137
+ {
138
+ "batch": 450,
139
+ "cv": 0.19790925203824197,
140
+ "cv_ema": 0.20349271362905041,
141
+ "recon": 1.1448991298675537
142
+ },
143
+ {
144
+ "batch": 500,
145
+ "cv": 0.20509912114315235,
146
+ "cv_ema": 0.20350877770419143,
147
+ "recon": 1.0240130424499512
148
+ },
149
+ {
150
+ "batch": 550,
151
+ "cv": 0.20451248072720773,
152
+ "cv_ema": 0.20351881473442157,
153
+ "recon": 1.0537513494491577
154
+ },
155
+ {
156
+ "batch": 600,
157
+ "cv": 0.184085603232447,
158
+ "cv_ema": 0.20332448261940184,
159
+ "recon": 1.098365306854248
160
+ },
161
+ {
162
+ "batch": 650,
163
+ "cv": 0.21316120274181655,
164
+ "cv_ema": 0.20342284982062597,
165
+ "recon": 1.03240966796875
166
+ },
167
+ {
168
+ "batch": 700,
169
+ "cv": 0.19525144262068833,
170
+ "cv_ema": 0.2033411357486266,
171
+ "recon": 0.9688878059387207
172
+ },
173
+ {
174
+ "batch": 750,
175
+ "cv": 0.19477563230793826,
176
+ "cv_ema": 0.20325548071421973,
177
+ "recon": 1.0756549835205078
178
+ },
179
+ {
180
+ "batch": 800,
181
+ "cv": 0.23088812783075377,
182
+ "cv_ema": 0.20353180718538505,
183
+ "recon": 0.9028558135032654
184
+ },
185
+ {
186
+ "batch": 850,
187
+ "cv": 0.19478954763463666,
188
+ "cv_ema": 0.20344438458987754,
189
+ "recon": 0.8116689920425415
190
+ },
191
+ {
192
+ "batch": 900,
193
+ "cv": 0.17564489737218603,
194
+ "cv_ema": 0.20316638971770062,
195
+ "recon": 0.8767406940460205
196
+ },
197
+ {
198
+ "batch": 950,
199
+ "cv": 0.20856704538854393,
200
+ "cv_ema": 0.20322039627440905,
201
+ "recon": 0.9646910429000854
202
+ }
203
+ ]
204
+ }