-
Notifications
You must be signed in to change notification settings - Fork 18
/
debug_precisions.yaml
66 lines (57 loc) · 2.54 KB
/
debug_precisions.yaml
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
# Copyright (c) Microsoft Corporation. All rights reserved.
# Licensed under a Microsoft Research License.
data:
devices: ['Pcat_Y81C76']
pretty_devices: ['Pcat-Pcat']
groups:
aR: [0]
aS: [0]
default_devices:
aR: 0
aS: 0
files: ['proc141006.csv']
signals: ["OD", "mRFP1", "EYFP", "ECFP"]
conditions: ["C6","C12"]
separate_conditions: true
params:
model: dr_constant_precisions
learning_boundaries: [250,1000]
learning_rates: [0.01,0.002,0.0002]
n_hidden_decoder_precisions: 5
constant: { init_x: 0.002, init_rfp: 0.0, init_yfp: 0.0, init_cfp: 0.0, init_luxR: 0.0, init_lasR: 0.0 }
# SHARED PARAMETERS
shared:
auto_precision_lognormal: { distribution: LogNormal, mu: -5.0, sigma: 2.0 }
a_precision_lognormal: { distribution: LogNormal, mu: 1.0, sigma: 2.0 }
dfp_precision_lognormal: { distribution: LogNormal, mu: -2.0, sigma: 1.5 }
global_conditioned:
conditioning: { devices: true, treatments: false }
global:
e76: { distribution: LogNormal, mu: -3.0, sigma: 1.0}
e81: { distribution: LogNormal, mu: -3.0, sigma: 1.0}
KGR_76: { distribution: LogNormal, mu: 2.0, sigma: 3.0}
KGR_81: { distribution: LogNormal, mu: -2.0, sigma: 3.0}
KGS_76: { distribution: LogNormal, mu: -2.0, sigma: 3.0}
KGS_81: { distribution: LogNormal, mu: 2.0, sigma: 3.0}
KR6: { distribution: LogNormal, mu: -6.0, sigma: 3.0}
KR12: { distribution: LogNormal, mu: -12.0, sigma: 3.0}
KS6: { distribution: LogNormal, mu: -12.0, sigma: 3.0}
KS12: { distribution: LogNormal, mu: -6.0, sigma: 3.0}
nR: { distribution: LogNormal, mu: 1.0, sigma: 0.25}
nS: { distribution: LogNormal, mu: 1.0, sigma: 0.25}
aYFP: { distribution: LogNormal, mu: 0.0, sigma: 2.0}
aCFP: { distribution: LogNormal, mu: 0.0, sigma: 2.0}
dR: { distribution: LogNormal, mu: -2.0, sigma: 1.0}
dS: { distribution: LogNormal, mu: -2.0, sigma: 1.0}
drfp: { distribution: dfp_precision_lognormal}
dyfp: { distribution: dfp_precision_lognormal}
dcfp: { distribution: dfp_precision_lognormal}
autoY: { distribution: auto_precision_lognormal}
autoC: { distribution: auto_precision_lognormal}
# LOCAL PARAMETERS:
local:
conditioning: { devices: true, treatments: False }
r: {distribution: LogNormal, mu: 0.0, sigma: 0.25}
K: {distribution: LogNormal, mu: 1.0, prec: 2.0}
tlag: {distribution: LogNormal, mu: 0.0, prec: 2.0}
rc: {distribution: LogNormal, mu: 0.0, sigma: 2.0} #blows up on 'prec', why?