Performance increase of broadening_methods and LDM optimizations.¶
Auto-download, calculate, and compare CO spectra calculated with several numerical approximation of the Voigt functions. We also show that the LDM optimization (“simple” or “min-RMS”) leads to a significantly faster computation for almost no impact on the Voigt computation.
from radis import SpectrumFactory, plot_diff
trunc_ref = 8
conditions = {
"molecule": "CO",
"wavenum_min": 1850,
"wavenum_max": 2325,
# "molecule": "H2O", #un-comment these lines to try with H2O
# "wavelength_min": 2400,
# "wavelength_max": 3000,
"path_length": 0.1,
"mole_fraction": 0.2,
"isotope": 1, # or "1,2" or "all"
"pressure": 1,
"wstep": 0.002,
# also measure interpolation time
"truncation": trunc_ref, # cm-1; Default value is 50 cm-1 but the value is decreased here to accelerate the execution of this example
"cutoff": 1e-27, # cm-1/(#.cm-2); Default is 1e-27
}
Tgas = 400
databank, database = "hitran", "default"
# databank, database = = "hitemp", "2010" #for H2O, on your machine (will take longer to compute the reference spectrum)
############## Reference spectrum - No optimization used ##############
conditions["optimization"] = None
## Using a convolution of Gaussian and Lorentzian, no optimization
conditions["broadening_method"] = (
"convolve" # Voigt = numpy.convolve("Gaussian", "Lorentzian")
)
sf = SpectrumFactory(**conditions)
sf.fetch_databank(databank)
s_conv = sf.eq_spectrum(Tgas=Tgas)
s_conv.name = f"Convolution : {s_conv.c['calculation_time']:.1f}s"
# Using a polynomial approximation, no optimization
conditions["broadening_method"] = (
"voigt_poly" # Voigt = polynomial approximation derived from Whithing
)
sf = SpectrumFactory(**conditions)
sf.fetch_databank(databank, database=database)
s_poly = sf.eq_spectrum(Tgas=Tgas)
s_poly.name = f"Poly. : {s_poly.c['calculation_time']:.1f}s"
# Compare the spectra
plot_diff(
s_conv,
s_poly,
var="absorbance",
yscale="log",
# yscale="linear",
)
############## Using LDM optimization ##############
"""
The objective of this section is to show how fast are the LDM optimizations.
The calculated spectra are compared to a reference (no optimization) where the
Voigts are calculated via numpy.convolve of a Gaussian and a Lorentzian.
**Conclusions:**
Using the 'linear' LDM optimization, it takes significantly less time to compute
the spectrum, whatever the broadening method employed. The relative difference is
always smaller than diff/ref < 1e-6. This difference can be improved by
changing from 'LDM = linear' to 'LDM = min-RMS' for a small cost in time.
"""
s_ref = s_conv
LDM_dic = {}
for method in ["convolve", "voigt_poly", "fft"]:
for LDM in ["simple", "min-RMS"]:
conditions["optimization"] = LDM
conditions["broadening_method"] = method
if method == "fft":
conditions["truncation"] = None
else:
conditions["truncation"] = s_ref.conditions[
"truncation"
] # cm-1 (Default value = 50)
sf = SpectrumFactory(**conditions)
sf.fetch_databank(databank)
s = sf.eq_spectrum(Tgas=Tgas)
s.name = f"{method} - {LDM}.: {s.c['calculation_time']:.1f} s"
LDM_dic[f"{LDM}_{method}"] = s
# Compare the spectrum to the reference
plot_diff(s_ref, s, "absorbance", yscale="log")
--------------------------------------------------------------------------------
CO - HITRAN - Downloading database
--------------------------------------------------------------------------------
Download:
- Downloading all isotopes for CO
Downloading isotopes: 0%| | 0/9 [00:00<?, ?it/s]
Downloading isotopes: 11%|█ | 1/9 [00:00<00:02, 2.84it/s]
Downloading isotopes: 22%|██▏ | 2/9 [00:00<00:02, 3.10it/s]
Downloading isotopes: 33%|███▎ | 3/9 [00:00<00:01, 3.17it/s]
Downloading isotopes: 44%|████▍ | 4/9 [00:01<00:01, 3.30it/s]
Downloading isotopes: 56%|█████▌ | 5/9 [00:01<00:01, 3.33it/s]
Downloading isotopes: 67%|██████▋ | 6/9 [00:01<00:00, 3.48it/s]
Downloading isotopes: 100%|██████████| 9/9 [00:01<00:00, 4.99it/s]
HITRAN database download complete
Added HITRAN-CO database in /home/docs/radis.json
2.00s - Loaded database
Calculating Equilibrium Spectrum
Physical Conditions
----------------------------------------
Tgas 400 K
isotope 1
medium air
mole_fraction 0.2
path_length 0.1 cm
pressure 1 bar
self_absorption True
species CO
state X
wavenum_max 2325.0000 cm-1
wavenum_min 1850.0000 cm-1
Computation Parameters
----------------------------------------
Tref 296 K
add_at_used
broadening_method convolve
cutoff 1e-27 cm-1/(#.cm-2)
dbformat hitran
dbpath
diluent air
folding_thresh 1e-06
include_neighbouring_lines True
isatom False
isneutral None
lbfunc None
memory_mapping_engine auto
neighbour_lines 0 cm-1
optimization None
parsum_mode full summation
pfsource default
potential_lowering None
pseudo_continuum_threshold 0
sparse_ldm True
truncation 8 cm-1
waveunit cm-1
wstep 0.002 cm-1
zero_padding 237501
----------------------------------------
2.00s - Spectrum calculated
--------------------------------------------------------------------------------
CO - HITRAN - Downloading database
--------------------------------------------------------------------------------
Download:
- All files already downloaded.
Caching to HDF5/H5 format:
- All files already cached.
0.03s - Loaded database
Calculating Equilibrium Spectrum
Physical Conditions
----------------------------------------
Tgas 400 K
isotope 1
medium air
mole_fraction 0.2
path_length 0.1 cm
pressure 1 bar
self_absorption True
species CO
state X
wavenum_max 2325.0000 cm-1
wavenum_min 1850.0000 cm-1
Computation Parameters
----------------------------------------
Tref 296 K
add_at_used
broadening_method voigt_poly
cutoff 1e-27 cm-1/(#.cm-2)
dbformat hitran
dbpath /home/docs/.radisdb/hitran/CO.h5
diluent air
folding_thresh 1e-06
include_neighbouring_lines True
isatom False
isneutral None
lbfunc None
memory_mapping_engine auto
neighbour_lines 0 cm-1
optimization None
parsum_mode full summation
pfsource default
potential_lowering None
pseudo_continuum_threshold 0
sparse_ldm True
truncation 8 cm-1
waveunit cm-1
wstep 0.002 cm-1
zero_padding 237501
----------------------------------------
0.14s - Spectrum calculated
--------------------------------------------------------------------------------
CO - HITRAN - Downloading database
--------------------------------------------------------------------------------
Download:
- All files already downloaded.
Caching to HDF5/H5 format:
- All files already cached.
0.03s - Loaded database
Calculating Equilibrium Spectrum
Physical Conditions
----------------------------------------
Tgas 400 K
isotope 1
medium air
mole_fraction 0.2
path_length 0.1 cm
pressure 1 bar
self_absorption True
species CO
state X
wavenum_max 2325.0000 cm-1
wavenum_min 1850.0000 cm-1
Computation Parameters
----------------------------------------
Tref 296 K
add_at_used numpy
broadening_method convolve
cutoff 1e-27 cm-1/(#.cm-2)
dbformat hitran
dbpath /home/docs/.radisdb/hitran/CO.h5
diluent air
folding_thresh 1e-06
include_neighbouring_lines True
isatom False
isneutral None
lbfunc None
memory_mapping_engine auto
neighbour_lines 0 cm-1
optimization simple
parsum_mode full summation
pfsource default
potential_lowering None
pseudo_continuum_threshold 0
sparse_ldm True
truncation 8 cm-1
waveunit cm-1
wstep 0.002 cm-1
zero_padding 237501
----------------------------------------
0.22s - Spectrum calculated
--------------------------------------------------------------------------------
CO - HITRAN - Downloading database
--------------------------------------------------------------------------------
Download:
- All files already downloaded.
Caching to HDF5/H5 format:
- All files already cached.
0.03s - Loaded database
Calculating Equilibrium Spectrum
Physical Conditions
----------------------------------------
Tgas 400 K
isotope 1
medium air
mole_fraction 0.2
path_length 0.1 cm
pressure 1 bar
self_absorption True
species CO
state X
wavenum_max 2325.0000 cm-1
wavenum_min 1850.0000 cm-1
Computation Parameters
----------------------------------------
Tref 296 K
add_at_used numpy
broadening_method convolve
cutoff 1e-27 cm-1/(#.cm-2)
dbformat hitran
dbpath /home/docs/.radisdb/hitran/CO.h5
diluent air
folding_thresh 1e-06
include_neighbouring_lines True
isatom False
isneutral None
lbfunc None
memory_mapping_engine auto
neighbour_lines 0 cm-1
optimization min-RMS
parsum_mode full summation
pfsource default
potential_lowering None
pseudo_continuum_threshold 0
sparse_ldm True
truncation 8 cm-1
waveunit cm-1
wstep 0.002 cm-1
zero_padding 237501
----------------------------------------
0.22s - Spectrum calculated
--------------------------------------------------------------------------------
CO - HITRAN - Downloading database
--------------------------------------------------------------------------------
Download:
- All files already downloaded.
Caching to HDF5/H5 format:
- All files already cached.
0.03s - Loaded database
Calculating Equilibrium Spectrum
Physical Conditions
----------------------------------------
Tgas 400 K
isotope 1
medium air
mole_fraction 0.2
path_length 0.1 cm
pressure 1 bar
self_absorption True
species CO
state X
wavenum_max 2325.0000 cm-1
wavenum_min 1850.0000 cm-1
Computation Parameters
----------------------------------------
Tref 296 K
add_at_used numpy
broadening_method voigt_poly
cutoff 1e-27 cm-1/(#.cm-2)
dbformat hitran
dbpath /home/docs/.radisdb/hitran/CO.h5
diluent air
folding_thresh 1e-06
include_neighbouring_lines True
isatom False
isneutral None
lbfunc None
memory_mapping_engine auto
neighbour_lines 0 cm-1
optimization simple
parsum_mode full summation
pfsource default
potential_lowering None
pseudo_continuum_threshold 0
sparse_ldm True
truncation 8 cm-1
waveunit cm-1
wstep 0.002 cm-1
zero_padding 237501
----------------------------------------
0.09s - Spectrum calculated
--------------------------------------------------------------------------------
CO - HITRAN - Downloading database
--------------------------------------------------------------------------------
Download:
- All files already downloaded.
Caching to HDF5/H5 format:
- All files already cached.
0.03s - Loaded database
Calculating Equilibrium Spectrum
Physical Conditions
----------------------------------------
Tgas 400 K
isotope 1
medium air
mole_fraction 0.2
path_length 0.1 cm
pressure 1 bar
self_absorption True
species CO
state X
wavenum_max 2325.0000 cm-1
wavenum_min 1850.0000 cm-1
Computation Parameters
----------------------------------------
Tref 296 K
add_at_used numpy
broadening_method voigt_poly
cutoff 1e-27 cm-1/(#.cm-2)
dbformat hitran
dbpath /home/docs/.radisdb/hitran/CO.h5
diluent air
folding_thresh 1e-06
include_neighbouring_lines True
isatom False
isneutral None
lbfunc None
memory_mapping_engine auto
neighbour_lines 0 cm-1
optimization min-RMS
parsum_mode full summation
pfsource default
potential_lowering None
pseudo_continuum_threshold 0
sparse_ldm True
truncation 8 cm-1
waveunit cm-1
wstep 0.002 cm-1
zero_padding 237501
----------------------------------------
0.10s - Spectrum calculated
--------------------------------------------------------------------------------
CO - HITRAN - Downloading database
--------------------------------------------------------------------------------
Download:
- All files already downloaded.
Caching to HDF5/H5 format:
- All files already cached.
0.03s - Loaded database
Calculating Equilibrium Spectrum
Physical Conditions
----------------------------------------
Tgas 400 K
isotope 1
medium air
mole_fraction 0.2
path_length 0.1 cm
pressure 1 bar
self_absorption True
species CO
state X
wavenum_max 2325.0000 cm-1
wavenum_min 1850.0000 cm-1
Computation Parameters
----------------------------------------
Tref 296 K
add_at_used numpy
broadening_method fft
cutoff 1e-27 cm-1/(#.cm-2)
dbformat hitran
dbpath /home/docs/.radisdb/hitran/CO.h5
diluent air
folding_thresh 1e-06
include_neighbouring_lines True
isatom False
isneutral None
lbfunc None
memory_mapping_engine auto
neighbour_lines 0 cm-1
optimization simple
parsum_mode full summation
pfsource default
potential_lowering None
pseudo_continuum_threshold 0
sparse_ldm True
truncation None cm-1
waveunit cm-1
wstep 0.002 cm-1
zero_padding 237501
----------------------------------------
1.60s - Spectrum calculated
--------------------------------------------------------------------------------
CO - HITRAN - Downloading database
--------------------------------------------------------------------------------
Download:
- All files already downloaded.
Caching to HDF5/H5 format:
- All files already cached.
0.03s - Loaded database
Calculating Equilibrium Spectrum
Physical Conditions
----------------------------------------
Tgas 400 K
isotope 1
medium air
mole_fraction 0.2
path_length 0.1 cm
pressure 1 bar
self_absorption True
species CO
state X
wavenum_max 2325.0000 cm-1
wavenum_min 1850.0000 cm-1
Computation Parameters
----------------------------------------
Tref 296 K
add_at_used numpy
broadening_method fft
cutoff 1e-27 cm-1/(#.cm-2)
dbformat hitran
dbpath /home/docs/.radisdb/hitran/CO.h5
diluent air
folding_thresh 1e-06
include_neighbouring_lines True
isatom False
isneutral None
lbfunc None
memory_mapping_engine auto
neighbour_lines 0 cm-1
optimization min-RMS
parsum_mode full summation
pfsource default
potential_lowering None
pseudo_continuum_threshold 0
sparse_ldm True
truncation None cm-1
waveunit cm-1
wstep 0.002 cm-1
zero_padding 237501
----------------------------------------
1.58s - Spectrum calculated
"""
Truncation speeds up spectrum computation. This section shows its effect on
performance for several LDM optimizations and broadening methods. For all
methods, computation time increases roughly linearly with truncation width,
while the residual versus the reference decreases exponentially.
For any broadening method, a residual of diff/ref < 1e-6 is reachable with
both LDM methods, with computation times from 0.1 to 2 seconds. This is much
faster than 'LDM = None' with 'broadening_method = convolve', which takes
about 20 seconds.
Note that the 'fft' method does not use truncation which explains the differences in the sides of the CO spectrum.
"""
s_ref = s_conv
import matplotlib.pyplot as plt
from tqdm import tqdm
from radis import get_residual
# Prepare the parameter grid for progress bar
trunc_values0 = [0.05, 0.1, 0.5, 1, 2, 3, 4, 5, 7]
LDM_values = [None, "simple", "min-RMS"]
method_values = ["convolve", "voigt_poly"]
conditions["verbose"] = False
total_iterations = (
sum(1 for trunc in trunc_values0 for LDM in LDM_values for method in method_values)
+ 2
)
progress = tqdm(total=total_iterations, desc="Processing")
results = {}
for LDM in LDM_values:
for method in method_values:
res_list, time_list = [], []
trunc_values = trunc_values0
if LDM is None and method == "convolve":
trunc_values = trunc_values0[:-1] # to avoid the longest computation
for trunc in trunc_values:
conditions["truncation"] = trunc
conditions["optimization"] = LDM
conditions["broadening_method"] = method
sf = SpectrumFactory(**conditions)
sf.fetch_databank(databank)
s = sf.eq_spectrum(Tgas=Tgas)
res_list.append(get_residual(s_ref, s, var="abscoeff"))
time_list.append(s.c["calculation_time"])
progress.update(1)
results[f"{LDM} - {method}"] = {
"trunc": trunc_values.copy(),
"residual": res_list,
"time": time_list,
}
## Add fft method (no truncation)
for LDM in ["simple", "min-RMS"]:
conditions["truncation"] = None
conditions["optimization"] = LDM
conditions["broadening_method"] = "fft"
sf = SpectrumFactory(**conditions)
sf.fetch_databank(databank)
s_fft = sf.eq_spectrum(Tgas=Tgas)
results[f"{LDM} - fft"] = {
"residual": get_residual(s_ref, s_fft, var="abscoeff"),
"time": s_fft.c["calculation_time"],
}
progress.update(1)
linestyles = {None: ":", "simple": "-", "min-RMS": "--"}
colors = {"convolve": "k", "voigt_poly": "r", "fft": "b"}
plt.figure("Time")
for LDM in LDM_values:
for method in method_values:
tag = f"{LDM} - {method}"
plt.plot(
results[tag]["trunc"],
results[tag]["time"],
f"{colors[method]}o",
linestyle=linestyles[LDM],
label=tag,
)
if LDM is not None:
tag = f"{LDM} - fft"
plt.plot(
max(trunc_values0) * 1.01,
results[tag]["time"],
f"{colors['fft']}*",
label=tag,
)
plt.xlabel("Truncation")
plt.ylabel("Computation time (s)")
plt.legend()
plt.figure("Residual")
for LDM in LDM_values:
for method, c in zip(method_values, colors):
tag = f"{LDM} - {method}"
plt.semilogy(
results[tag]["trunc"],
results[tag]["residual"],
f"{colors[method]}o",
linestyle=linestyles[LDM],
label=tag,
)
if LDM is not None:
tag = f"{LDM} - fft"
plt.plot(
max(trunc_values0) * 1.01,
results[tag]["residual"],
f"{colors['fft']}*",
label=tag,
)
plt.xlabel("Truncation")
plt.ylabel("Residual (abscoeff)")
plt.legend()
Processing: 0%| | 0/56 [00:00<?, ?it/s]0.03s - Loaded database
/home/docs/checkouts/readthedocs.org/user_builds/radis/checkouts/latest/radis/misc/warning.py:436: CollisionalBroadeningWarning: Large error (57.8%) in pressure broadening. Increase broadening width / reduce wstep. Use .plot_broadening() to visualize each line broadening
warnings.warn(WarningType(message))
0.03s - Loaded database
/home/docs/checkouts/readthedocs.org/user_builds/radis/checkouts/latest/radis/misc/warning.py:436: CollisionalBroadeningWarning: Large error (36.3%) in pressure broadening. Increase broadening width / reduce wstep. Use .plot_broadening() to visualize each line broadening
warnings.warn(WarningType(message))
Processing: 4%|▎ | 2/56 [00:00<00:03, 16.67it/s]0.03s - Loaded database
/home/docs/checkouts/readthedocs.org/user_builds/radis/checkouts/latest/radis/misc/warning.py:436: CollisionalBroadeningWarning: Large error (8.1%) in pressure broadening. Increase broadening width / reduce wstep. Use .plot_broadening() to visualize each line broadening
warnings.warn(WarningType(message))
0.03s - Loaded database
/home/docs/checkouts/readthedocs.org/user_builds/radis/checkouts/latest/radis/misc/warning.py:436: CollisionalBroadeningWarning: Large error (4.1%) in pressure broadening. Increase broadening width / reduce wstep. Use .plot_broadening() to visualize each line broadening
warnings.warn(WarningType(message))
Processing: 7%|▋ | 4/56 [00:00<00:03, 13.79it/s]0.03s - Loaded database
/home/docs/checkouts/readthedocs.org/user_builds/radis/checkouts/latest/radis/misc/warning.py:436: CollisionalBroadeningWarning: Large error (2.0%) in pressure broadening. Increase broadening width / reduce wstep. Use .plot_broadening() to visualize each line broadening
warnings.warn(WarningType(message))
0.03s - Loaded database
/home/docs/checkouts/readthedocs.org/user_builds/radis/checkouts/latest/radis/misc/warning.py:436: CollisionalBroadeningWarning: Large error (1.4%) in pressure broadening. Increase broadening width / reduce wstep. Use .plot_broadening() to visualize each line broadening
warnings.warn(WarningType(message))
Processing: 11%|█ | 6/56 [00:00<00:07, 6.40it/s]0.03s - Loaded database
/home/docs/checkouts/readthedocs.org/user_builds/radis/checkouts/latest/radis/misc/warning.py:436: CollisionalBroadeningWarning: Large error (1.0%) in pressure broadening. Increase broadening width / reduce wstep. Use .plot_broadening() to visualize each line broadening
warnings.warn(WarningType(message))
Processing: 12%|█▎ | 7/56 [00:01<00:12, 3.90it/s]0.03s - Loaded database
Processing: 14%|█▍ | 8/56 [00:02<00:19, 2.41it/s]0.03s - Loaded database
0.03s - Loaded database
Processing: 18%|█▊ | 10/56 [00:02<00:12, 3.81it/s]0.03s - Loaded database
0.03s - Loaded database
Processing: 21%|██▏ | 12/56 [00:02<00:08, 5.32it/s]0.03s - Loaded database
0.03s - Loaded database
Processing: 25%|██▌ | 14/56 [00:02<00:06, 6.51it/s]0.03s - Loaded database
0.03s - Loaded database
Processing: 29%|██▊ | 16/56 [00:02<00:05, 6.94it/s]0.03s - Loaded database
0.03s - Loaded database
Processing: 32%|███▏ | 18/56 [00:03<00:05, 7.02it/s]0.03s - Loaded database
0.03s - Loaded database
Processing: 36%|███▌ | 20/56 [00:03<00:04, 8.10it/s]0.03s - Loaded database
0.03s - Loaded database
Processing: 39%|███▉ | 22/56 [00:03<00:04, 8.47it/s]0.03s - Loaded database
Processing: 41%|████ | 23/56 [00:03<00:03, 8.31it/s]0.03s - Loaded database
Processing: 43%|████▎ | 24/56 [00:03<00:04, 7.96it/s]0.03s - Loaded database
Processing: 45%|████▍ | 25/56 [00:03<00:04, 7.40it/s]0.03s - Loaded database
Processing: 46%|████▋ | 26/56 [00:04<00:04, 6.35it/s]0.03s - Loaded database
0.03s - Loaded database
Processing: 50%|█████ | 28/56 [00:04<00:03, 7.93it/s]0.03s - Loaded database
0.03s - Loaded database
Processing: 54%|█████▎ | 30/56 [00:04<00:02, 8.81it/s]0.03s - Loaded database
Processing: 55%|█████▌ | 31/56 [00:04<00:02, 8.85it/s]0.03s - Loaded database
Processing: 57%|█████▋ | 32/56 [00:04<00:02, 8.77it/s]0.03s - Loaded database
Processing: 59%|█████▉ | 33/56 [00:04<00:02, 8.70it/s]0.03s - Loaded database
Processing: 61%|██████ | 34/56 [00:05<00:02, 8.56it/s]0.03s - Loaded database
Processing: 62%|██████▎ | 35/56 [00:05<00:02, 8.26it/s]0.03s - Loaded database
0.03s - Loaded database
Processing: 66%|██████▌ | 37/56 [00:05<00:01, 9.74it/s]0.03s - Loaded database
0.03s - Loaded database
Processing: 70%|██████▉ | 39/56 [00:05<00:01, 10.21it/s]0.03s - Loaded database
0.03s - Loaded database
Processing: 73%|███████▎ | 41/56 [00:05<00:01, 9.35it/s]0.03s - Loaded database
Processing: 75%|███████▌ | 42/56 [00:05<00:01, 8.71it/s]0.03s - Loaded database
Processing: 77%|███████▋ | 43/56 [00:06<00:01, 7.87it/s]0.03s - Loaded database
Processing: 79%|███████▊ | 44/56 [00:06<00:01, 6.63it/s]0.03s - Loaded database
0.03s - Loaded database
Processing: 82%|████████▏ | 46/56 [00:06<00:01, 8.17it/s]0.03s - Loaded database
0.03s - Loaded database
Processing: 86%|████████▌ | 48/56 [00:06<00:00, 8.93it/s]0.03s - Loaded database
Processing: 88%|████████▊ | 49/56 [00:06<00:00, 8.94it/s]0.03s - Loaded database
Processing: 89%|████████▉ | 50/56 [00:06<00:00, 8.86it/s]0.03s - Loaded database
Processing: 91%|█████████ | 51/56 [00:06<00:00, 8.74it/s]0.03s - Loaded database
Processing: 93%|█████████▎| 52/56 [00:07<00:00, 8.56it/s]0.03s - Loaded database
Processing: 95%|█████████▍| 53/56 [00:07<00:00, 8.24it/s]0.03s - Loaded database
Processing: 96%|█████████▋| 54/56 [00:08<00:01, 1.88it/s]0.03s - Loaded database
Processing: 98%|█████████▊| 55/56 [00:10<00:00, 1.19it/s]
<matplotlib.legend.Legend object at 0x73c595758980>
Total running time of the script: (0 minutes 32.643 seconds)