AbstractPhil commited on
Commit
08d27f6
·
verified ·
1 Parent(s): dbef1ac

Ablation D-MID-D1_cosine

Browse files
D/MID/D1_cosine/seed0/final_report.json ADDED
@@ -0,0 +1,203 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "config": {
3
+ "group": "D",
4
+ "variant": "D1_cosine",
5
+ "band": "MID",
6
+ "seed": 0,
7
+ "phase": 1,
8
+ "overrides": {
9
+ "scheduler": "cosine"
10
+ },
11
+ "description": "D-MID-D1_cosine"
12
+ },
13
+ "run_config": {
14
+ "matrix_v": 64,
15
+ "D": 8,
16
+ "patch_size": 16,
17
+ "hidden": 64,
18
+ "depth": 1,
19
+ "n_cross_layers": 1,
20
+ "n_heads": 4,
21
+ "max_alpha": 0.2,
22
+ "alpha_init": -2.0,
23
+ "img_size": 64,
24
+ "batch_size": 128,
25
+ "lr": 0.0001,
26
+ "epochs": 1,
27
+ "weight_decay": 0.0,
28
+ "use_cv_ema": true,
29
+ "cv_ema_alpha": 0.01,
30
+ "cv_alignment_epochs": 0,
31
+ "cv_measure_every": 50,
32
+ "cv_sigma_scale": 0.3,
33
+ "boost": 0.5,
34
+ "cross_attn_clip": 0.5,
35
+ "allowed_types": [
36
+ 0,
37
+ 1,
38
+ 2,
39
+ 3,
40
+ 4,
41
+ 5,
42
+ 6,
43
+ 7,
44
+ 8,
45
+ 9,
46
+ 10,
47
+ 11,
48
+ 12,
49
+ 13,
50
+ 14,
51
+ 15
52
+ ],
53
+ "train_size": 1000000,
54
+ "val_size": 10000,
55
+ "num_workers": 2,
56
+ "report_every": 100,
57
+ "major_report_every": 10,
58
+ "save_every": 5,
59
+ "seed": 0,
60
+ "hf_repo": "AbstractPhil/geolip-svae-batteries",
61
+ "upload": false
62
+ },
63
+ "cv_ema_final": 0.3800111213601007,
64
+ "cv_last": 0.3350629158345097,
65
+ "predicted_band": "MID",
66
+ "expected_band": "MID",
67
+ "band_match": true,
68
+ "test_mse": 1.0095101594924927,
69
+ "recon_ema": 1.0121322766540295,
70
+ "S0": 3.462233781814575,
71
+ "SD": 2.1595096588134766,
72
+ "ratio": 1.6032499561425297,
73
+ "erank": 7.909370422363281,
74
+ "observed_sphere_cv": 0.352242005946348,
75
+ "uniform_sphere_cv_prediction": 0.35679185069620456,
76
+ "band_deviation": -0.004549844749856535,
77
+ "params_count": 183339,
78
+ "wallclock_seconds": 33.60575604438782,
79
+ "batches_completed": 1000,
80
+ "batch_limit": 1000,
81
+ "cv_trajectory": [
82
+ {
83
+ "batch": 0,
84
+ "cv": 0.3849132290425209,
85
+ "cv_ema": 0.3849132290425209,
86
+ "recon": 1.9087696075439453
87
+ },
88
+ {
89
+ "batch": 50,
90
+ "cv": 0.363281001610386,
91
+ "cv_ema": 0.38469690676819956,
92
+ "recon": 1.6203091144561768
93
+ },
94
+ {
95
+ "batch": 100,
96
+ "cv": 0.35034573090098703,
97
+ "cv_ema": 0.3843533950095274,
98
+ "recon": 1.3560752868652344
99
+ },
100
+ {
101
+ "batch": 150,
102
+ "cv": 0.3778979617480635,
103
+ "cv_ema": 0.38428884067691277,
104
+ "recon": 1.3715566396713257
105
+ },
106
+ {
107
+ "batch": 200,
108
+ "cv": 0.3555635009013425,
109
+ "cv_ema": 0.38400158727915706,
110
+ "recon": 1.131544589996338
111
+ },
112
+ {
113
+ "batch": 250,
114
+ "cv": 0.39336818424440073,
115
+ "cv_ema": 0.3840952532488095,
116
+ "recon": 1.343937873840332
117
+ },
118
+ {
119
+ "batch": 300,
120
+ "cv": 0.38902556372899255,
121
+ "cv_ema": 0.38414455635361133,
122
+ "recon": 1.0601253509521484
123
+ },
124
+ {
125
+ "batch": 350,
126
+ "cv": 0.3266918388693671,
127
+ "cv_ema": 0.3835700291787689,
128
+ "recon": 1.004065752029419
129
+ },
130
+ {
131
+ "batch": 400,
132
+ "cv": 0.35717812100897806,
133
+ "cv_ema": 0.383306110097071,
134
+ "recon": 1.0957077741622925
135
+ },
136
+ {
137
+ "batch": 450,
138
+ "cv": 0.35888517594145053,
139
+ "cv_ema": 0.3830619007555148,
140
+ "recon": 1.0435925722122192
141
+ },
142
+ {
143
+ "batch": 500,
144
+ "cv": 0.3550559605676852,
145
+ "cv_ema": 0.38278184135363647,
146
+ "recon": 0.9685532450675964
147
+ },
148
+ {
149
+ "batch": 550,
150
+ "cv": 0.35859162010067197,
151
+ "cv_ema": 0.38253993914110684,
152
+ "recon": 1.0460222959518433
153
+ },
154
+ {
155
+ "batch": 600,
156
+ "cv": 0.3553032277454505,
157
+ "cv_ema": 0.3822675720271503,
158
+ "recon": 0.9279849529266357
159
+ },
160
+ {
161
+ "batch": 650,
162
+ "cv": 0.3465789004456751,
163
+ "cv_ema": 0.38191068531133554,
164
+ "recon": 1.2364634275436401
165
+ },
166
+ {
167
+ "batch": 700,
168
+ "cv": 0.38431601092236467,
169
+ "cv_ema": 0.3819347385674458,
170
+ "recon": 1.0680174827575684
171
+ },
172
+ {
173
+ "batch": 750,
174
+ "cv": 0.3212262994490936,
175
+ "cv_ema": 0.3813276541762623,
176
+ "recon": 1.073204755783081
177
+ },
178
+ {
179
+ "batch": 800,
180
+ "cv": 0.3596286476578066,
181
+ "cv_ema": 0.38111066411107775,
182
+ "recon": 1.079711675643921
183
+ },
184
+ {
185
+ "batch": 850,
186
+ "cv": 0.34325553663791186,
187
+ "cv_ema": 0.38073211283634606,
188
+ "recon": 1.1631325483322144
189
+ },
190
+ {
191
+ "batch": 900,
192
+ "cv": 0.3540351930154363,
193
+ "cv_ema": 0.38046514363813694,
194
+ "recon": 0.9824771285057068
195
+ },
196
+ {
197
+ "batch": 950,
198
+ "cv": 0.3350629158345097,
199
+ "cv_ema": 0.3800111213601007,
200
+ "recon": 1.175780177116394
201
+ }
202
+ ]
203
+ }