-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathsynthetic_comparisons_Frobenius.py
185 lines (158 loc) · 6.75 KB
/
synthetic_comparisons_Frobenius.py
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
import numpy as np
from matplotlib import pyplot as plt
import NMF_Frobenius as nmf_f
from nn_fac.nmf import nmf
#import tensorly as tl #perso branch
#from tensorly.decomposition import non_negative_parafac_hals
import pandas as pd
import plotly.express as px
import plotly.graph_objects as go
import sys
import plotly.io as pio
import time
pio.kaleido.scope.mathjax = None
# Personnal comparison toolbox
# you can get it at
# https://github.com/cohenjer/shootout
from shootout.methods.runners import run_and_track
from shootout.methods.plotters import plot_speed_comparison
plt.close('all')
# --------------------- Choose parameters for grid tests ------------ #
if len(sys.argv)==1 or int(sys.argv[1])==0:
seeds = [] #no run
skip=True
else:
seeds = list(np.arange(int(sys.argv[1])))
skip=False
variables = {
"add_track" : {"distribution" : "uniform"},
"mnr" : [[200,100,5],[1000,400,20]],
"NbIter" : [200], # for Lee and Seung also
"NbIter_inner" : 100,
"SNR" : [100],
"delta" : 0.1,
"seed" : seeds,
"distribution" : "uniform",
"show_it" : 100,
"epsilon" : 1e-8,
"tol" : 0
}
name = "l2_run-10-05-2023"
algs = ["MU_Fro","fastMU_Fro_ex","GD_Fro", "NeNMF_Fro", "HALS", "fastMU_Fro"]
@run_and_track(algorithm_names=algs, path_store="Results/", name_store=name,
skip=skip, **variables)
def one_run(**cfg):
m, n, r = cfg["mnr"]
# Fixed the signal
rng = np.random.RandomState(cfg["seed"]+20)
Worig = rng.rand(m, r)
Horig = rng.rand(r, n)
Vorig = Worig.dot(Horig)
# prints
verbose = True
# Initialization for H0 as a random matrix
Hini = rng.rand(r, n)
Wini = rng.rand(m, r) #sparse.random(rV, cW, density=0.25).toarray()
# adding Gaussian noise to the observed data
N = rng.randn(m,n)
sigma = 10**(-cfg["SNR"]/20)*np.linalg.norm(Vorig)/np.linalg.norm(N)
V = Vorig + sigma*N
# One noise, one init; NMF is not unique and nncvx so we will find several results
error0, W0, H0, toc0, cnt0 = nmf_f.NMF_Lee_Seung(V, Wini, Hini, cfg["NbIter"], cfg["NbIter_inner"],tol=cfg["tol"], legacy=False, delta=cfg["delta"], verbose=verbose)
error2, W2, H2, toc2, cnt2 = nmf_f.Grad_descent(V , Wini, Hini, cfg["NbIter"], cfg["NbIter_inner"], tol=cfg["tol"], delta=cfg["delta"], verbose=verbose)
error3, W3, H3, toc3, cnt3 = nmf_f.NeNMF(V, Wini, Hini, tol=cfg["tol"], nb_inner=cfg["NbIter_inner"], itermax=cfg["NbIter"], delta=cfg["delta"], verbose=verbose)
# Fewer max iter because too slow
# With Tensorly --> too slow
# callback def
#toc4=[]
#error4=[]
#cnt4=[]
#norm_tensor = tl.norm(V,2)**2
#def callback_call(cp_tensor,error, inner_iter=None):
#toc4.append(time.perf_counter())
#error4.append(np.sqrt(error*2*norm_tensor)/m/n) #err in cp is 1/2 \| \|_F^2
#if inner_iter is not None:
#cnt4.append(inner_iter)
#[W4, H4], _, _ = non_negative_parafac_hals(V, r, init=(None,[np.copy(Wini),np.copy(Hini).T]), n_iter_max=cfg["NbIter"], tol=cfg["tol"], return_errors=True, inner_iter_max=cfg["NbIter_inner"], inner_tol=cfg["delta"]*5, verbose=verbose, callback=callback_call)
#toc4 = [toc4[i]-toc4[0] for i in range(len(toc4))]
# Q: error4 is right format?
_, _, error4, toc4, cnt4 = nmf(V, r, init="random", n_iter_max=cfg["NbIter"], tol=cfg["tol"], return_costs=True, NbIter_inner=cfg["NbIter_inner"], delta=cfg["delta"], verbose=verbose)
error1, W1, H1, toc1, cnt1 = nmf_f.NeNMF_optimMajo(V, Wini, Hini, tol=cfg["tol"], itermax=cfg["NbIter"], nb_inner=cfg["NbIter_inner"], delta=cfg["delta"], verbose=verbose, use_best=False, gamma=1)
error5, W5, H5, toc5, cnt5 = nmf_f.NMF_proposed_Frobenius(V, Wini, Hini, cfg["NbIter"], cfg["NbIter_inner"], tol=cfg["tol"], delta=cfg["delta"], verbose=verbose, gamma=1.9)
# algs = ["MU_Fro","fastMU_Fro_ex","GD_Fro", "NeNMF_Fro", "HALS", "fastMU_Fro"]
return {"errors" : [error0, error1, error2, error3, error4, error5],
"timings" : [toc0, toc1, toc2, toc3, toc4, toc5],
"cnt" : [cnt0[::10], cnt1[::10], cnt2[::10], cnt3[::10], cnt4[::10], cnt5[::10]]
}
# -------------------- Post-Processing ------------------- #
import pandas as pd
import shootout.methods.post_processors as pp
pio.templates.default= "plotly_white"
df = pd.read_pickle("Results/"+name)
# Remove extrapolation
df = df[df["algorithm"] != "fastMU_Fro_ex"]
# TODO: shootout plots
## Using shootout for plotting and postprocessing
#thresh = np.logspace(-3,-8,50)
#scores_time, scores_it, timings, iterations = pp.find_best_at_all_thresh(df,thresh, nb_seeds)
# ----------------------- Plot --------------------------- #
#fig_winner = plot_speed_comparison(thresh, scores_time, scores_it, legend=algs)
#fig_winner.show()
# Adding in results errors at specific timings and iterations
#df = pp.error_at_time_or_it(df, time_stamps=[0.1, 0.5, 1], it_stamps=[10, 50, 300])
# Group up columns
#df = pp.regroup_columns(df, keys=["mnr"], how_many=3)
# Interpolating time (choose fewer points for better vis), adaptive grid since time varies across plots
ovars_interp = ["mnr", "SNR", "algorithm"]
df = pp.interpolate_time_and_error(df, npoints = 100, adaptive_grid=True, groups=ovars_interp)
# Making a convergence plot dataframe
# We will show convergence plots for various sigma values, with only n=100
ovars = ["mnr", "SNR","seed"]
df_conv = pp.df_to_convergence_df(df, groups=True, groups_names=ovars, other_names=ovars, err_name="errors_interp", time_name="timings_interp")
df_conv = df_conv.rename(columns={"timings_interp": "timings", "errors_interp": "errors"})
# Median plot
df_conv_median_time = pp.median_convergence_plot(df_conv, type_x="timings", mean=False)
# Convergence plots with all runs
pxfig = px.line(
df_conv_median_time,
x="timings",
y= "errors",
color='algorithm',
line_dash='algorithm',
facet_col="mnr",
log_y=True,
facet_col_spacing=0.1,
#log_x=True,
#error_y="q_errors_p",
#error_y_minus="q_errors_m",
)
# Final touch
pxfig.update_traces(
selector=dict(),
line_width=2.5,
#error_y_thickness = 0.3,
)
pxfig.update_layout(
font_size = 12,
width=450*1.62, # in px
height=450,
xaxis1=dict(range=[0,3],title_text="Time (s)"),
xaxis2=dict(range=[0,0.2],title_text="Time (s)"),
yaxis1=dict(title_text="Fit"),
yaxis2=dict(title_text="")
)
pxfig.update_xaxes(
matches = None,
showticklabels = True
)
pxfig.update_yaxes(
matches=None,
showticklabels=True
)
# updating titles
for i,ann in enumerate(pxfig.layout.annotations):
if ann.text[:3]=="mnr":
ann.text="[m,n,r]"+ann.text[3:]
pxfig.write_image("Results/"+name+".pdf")
pxfig.write_image("Results/"+name+".pdf")
pxfig.show()