MOPS magnitude estimator using a Convolutional Neural Network¶
Requirements for this notebook¶
This code tested on version numbers shown
python >=3.7.12 numpy >= 1.19.5 scipy >= 1.7.3 keras >= 2.6.0 tensorflow >= 2.6.2 matplotlib >= 3.5.1
Not using dropout¶
Link about how dropout was responsible for the bias in the output fluxes:
This is useful reading!
In [1]:
Copied!
import os, time, sys, pickle, glob, gc
"""Import the basics: numpy, matplotlib, etc."""
import numpy as np
from scipy import interpolate as interp
import matplotlib.pyplot as pyl
import matplotlib.gridspec as gridspec
import pickle
"""Import keras and other ML tools"""
import tensorflow as tf
import keras
import keras.backend as backend
from ensemble_magregressor import *
import os, time, sys, pickle, glob, gc
"""Import the basics: numpy, matplotlib, etc."""
import numpy as np
from scipy import interpolate as interp
import matplotlib.pyplot as pyl
import matplotlib.gridspec as gridspec
import pickle
"""Import keras and other ML tools"""
import tensorflow as tf
import keras
import keras.backend as backend
from ensemble_magregressor import *
Download the data files¶
If you wish to train a new network based on the data used to train the networks published in Chapter 9, please set download_data = True, and execute the following cell. This will download roughly 3.2 gb of tarballs and expand.
In [8]:
Copied!
warps_url = "https://www.canfar.net/storage/vault/file/AstroDataCitationDOI/CISTI.CANFAR/24.0084/data/"
## set this to true to download the training data files
download_data = False
if not os.path.isfile('warps_data.tgz') and download_data:
## use curl
#os.system(f'curl -o warps_data.tgz {warps_url}/warps_data.tgz')
## use wget
os.system(f'wget {warps_url}/warps_data.tgz')
#os.system('tar -xvzf warps_data.tgz')
if not os.path.isfile('kbmod_results_data.tgz') and download_data:
## use curl
#os.system(f'curl -o kbmod_results_data.tgz {warps_url}/kbmod_results_data.tgz')
##use wget
os.system(f'wget {warps_url}/kbmod_results_data.tgz')
os.system('tar -xvzf kbmod_results_data.tgz')
warps_url = "https://www.canfar.net/storage/vault/file/AstroDataCitationDOI/CISTI.CANFAR/24.0084/data/"
## set this to true to download the training data files
download_data = False
if not os.path.isfile('warps_data.tgz') and download_data:
## use curl
#os.system(f'curl -o warps_data.tgz {warps_url}/warps_data.tgz')
## use wget
os.system(f'wget {warps_url}/warps_data.tgz')
#os.system('tar -xvzf warps_data.tgz')
if not os.path.isfile('kbmod_results_data.tgz') and download_data:
## use curl
#os.system(f'curl -o kbmod_results_data.tgz {warps_url}/kbmod_results_data.tgz')
##use wget
os.system(f'wget {warps_url}/kbmod_results_data.tgz')
os.system('tar -xvzf kbmod_results_data.tgz')
Setup some hyper parameters and some variables for input data.
In [2]:
Copied!
visits = ['2022-08-01-AS1_July', #r
'2022-08-22-AS2', #r
'2022-08-23-AS2', #r
'2022-08-26-AS2', #r
]
save_model_path = 'MyNewModel'
print(f'Will save the trained model to {save_model_path}')
learning_rate = 0.00025
num_dense_nodes = 8
num_dense_layers = 2
num_filters = 6
train_epochs = 250
num_models = 5
## if you find that your model is reporting nan's for loss during training, reduce the batch_size. 768 seems to work well for the reallysmall model
batch_size = 4096 #768
test_fraction = 0.05
useMedForNans = False #otherwise zero is used
image_data_type = 'float32'
double_flip = True # mirror vertically and horizontally augmentation
shuffle_augment = True
rotate_augment = True
gridType = '_tg'
## Keep sources within the range of brightness of interest. Feel free to explore
nukeFaint = 26.5
nukeBright = 21.
dist_lim = 3.0 #association to planted source
rate_diff_lim = 45.0 #pix per day
save_model_iteration = False
save_model = False
useSampleWeights = True
warps_path = f'./warps'
chips = []
for i in range(40):
chips.append(str(i).zfill(2))
## setup an array that rotates all sources to have the same RA/Dec orientation. This may or may not be important
## the index is the # of 90 degree rotations applied to the stack
rots = np.array([2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 2, 0, 0])
visits = ['2022-08-01-AS1_July', #r
'2022-08-22-AS2', #r
'2022-08-23-AS2', #r
'2022-08-26-AS2', #r
]
save_model_path = 'MyNewModel'
print(f'Will save the trained model to {save_model_path}')
learning_rate = 0.00025
num_dense_nodes = 8
num_dense_layers = 2
num_filters = 6
train_epochs = 250
num_models = 5
## if you find that your model is reporting nan's for loss during training, reduce the batch_size. 768 seems to work well for the reallysmall model
batch_size = 4096 #768
test_fraction = 0.05
useMedForNans = False #otherwise zero is used
image_data_type = 'float32'
double_flip = True # mirror vertically and horizontally augmentation
shuffle_augment = True
rotate_augment = True
gridType = '_tg'
## Keep sources within the range of brightness of interest. Feel free to explore
nukeFaint = 26.5
nukeBright = 21.
dist_lim = 3.0 #association to planted source
rate_diff_lim = 45.0 #pix per day
save_model_iteration = False
save_model = False
useSampleWeights = True
warps_path = f'./warps'
chips = []
for i in range(40):
chips.append(str(i).zfill(2))
## setup an array that rotates all sources to have the same RA/Dec orientation. This may or may not be important
## the index is the # of 90 degree rotations applied to the stack
rots = np.array([2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 2, 0, 0])
Will save the trained model to MyNewModel
Read in the data.¶
Input data include the plantList files which record the magnitudes of injected artificial sources, and stamps files which are the shift-stacks of each candidate source. Zeropoints are used to convert magnitudes to instrumental
In [3]:
Copied!
with open('zeropoints.dat') as han:
data = han.readlines()
zpos = {}
for i in range(len(data)):
(v,c,z) = data[i].split()
if v not in list(zpos.keys()):
zpos[v] = {}
zpos[v][c] = float(z)
stamp_files, fs, kb_xys, zeropoints = [], [], [], []
counter = 0
for j, v in enumerate(visits):
for i, c in enumerate(chips):
stamps_path = f'{warps_path}/{v}'
plantLists_path = f'{warps_path}/{v}/'
kbmod_results_path = f'./kbmod_results/{v}/results_{c}/'
stamp_files.append(f'{stamps_path}/stamps{gridType}_{c}_w_sr.pickle')
zpo = zpos[v][c]
### load the kbmod results
kb_xy = []
with open(f'{kbmod_results_path}/results_.txt') as han:
data = han.readlines()
for ii in range(len(data)):
s = data[ii].split()
x, y = float(s[5]), float(s[7])
repeat = False
for jj in range(len(kb_xy)):
if kb_xy[jj][0]==x and kb_xy[jj][1]==y:
repeat = True
break
if not repeat:
kb_xy.append([float(s[5]) , float(s[7]) , float(s[9]), float(s[11]), float(s[1]), 0.0])
kb_xy = np.array(kb_xy)
### load the plantlist sources
p_xy = []
plantLists = glob.glob(f'{plantLists_path}/{c}/*plantList')
plantLists.sort()
with open(plantLists[0]) as han:
data = han.readlines()
for ii in range(1,len(data)):
s = data[ii].split()
x,y = float(s[3]), float(s[4])
rate = float(s[5])*24./0.187
repeat = False
for jj in range(len(p_xy)):
if p_xy[jj][0] == x and p_xy[jj][1]==y:
p_xy[jj][2]-=0.75
repeat = True
if not repeat:
p_xy.append([x, y, float(s[9]), rate, 0])
if len(p_xy)>0:
p_xy = np.array(p_xy)
p_xy = p_xy[np.argsort(p_xy[:,2])]
p_xy = p_xy[np.where((p_xy[:,2]>nukeBright)&(p_xy[:,2]<nukeFaint))]
#label the good and bad sources
for ii in range(len(p_xy)):
d = ((p_xy[ii, 0] - kb_xy[:, 0])**2 + (p_xy[ii, 1] - kb_xy[:, 1])**2 )**0.5
d_rate = (p_xy[ii, 3] - (kb_xy[:, 2]**2 + kb_xy[:, 3]**2)**0.5)
ww = np.where((d<dist_lim)&(np.abs(d_rate)<rate_diff_lim))
if len(ww[0])>0:
for arg in ww[0]:
kb_xy[arg,-1] = p_xy[ii,2]-zpo
w_good = np.where(kb_xy[:,-1]!=0)
#load the stamps
with open(stamp_files[-1], 'rb') as han:
f = pickle.load(han)
if rots[counter%len(chips)]!=0:# and rots[counter%len(chips)]!=2:
f = np.rot90(f, k=-rots[counter%len(chips)], axes=(1, 2))
counter+=1
### clip to avoid the crazy min pixel values
f = np.clip(f, -3500., np.max(f))
f_med = np.nanmedian(f)
f = f[w_good]
kb_xy = kb_xy[w_good]
fs.append(f)
kb_xys.append(kb_xy)
zeropoints.append(np.zeros(len(f), dtype='float64')+zpo)
print(v, c,f.shape)
sns_frames = np.concatenate(fs)
kb_xys = np.concatenate(kb_xys)
zeropoints = np.concatenate(zeropoints)
print('Total sources:', zeropoints.shape)
with open('zeropoints.dat') as han:
data = han.readlines()
zpos = {}
for i in range(len(data)):
(v,c,z) = data[i].split()
if v not in list(zpos.keys()):
zpos[v] = {}
zpos[v][c] = float(z)
stamp_files, fs, kb_xys, zeropoints = [], [], [], []
counter = 0
for j, v in enumerate(visits):
for i, c in enumerate(chips):
stamps_path = f'{warps_path}/{v}'
plantLists_path = f'{warps_path}/{v}/'
kbmod_results_path = f'./kbmod_results/{v}/results_{c}/'
stamp_files.append(f'{stamps_path}/stamps{gridType}_{c}_w_sr.pickle')
zpo = zpos[v][c]
### load the kbmod results
kb_xy = []
with open(f'{kbmod_results_path}/results_.txt') as han:
data = han.readlines()
for ii in range(len(data)):
s = data[ii].split()
x, y = float(s[5]), float(s[7])
repeat = False
for jj in range(len(kb_xy)):
if kb_xy[jj][0]==x and kb_xy[jj][1]==y:
repeat = True
break
if not repeat:
kb_xy.append([float(s[5]) , float(s[7]) , float(s[9]), float(s[11]), float(s[1]), 0.0])
kb_xy = np.array(kb_xy)
### load the plantlist sources
p_xy = []
plantLists = glob.glob(f'{plantLists_path}/{c}/*plantList')
plantLists.sort()
with open(plantLists[0]) as han:
data = han.readlines()
for ii in range(1,len(data)):
s = data[ii].split()
x,y = float(s[3]), float(s[4])
rate = float(s[5])*24./0.187
repeat = False
for jj in range(len(p_xy)):
if p_xy[jj][0] == x and p_xy[jj][1]==y:
p_xy[jj][2]-=0.75
repeat = True
if not repeat:
p_xy.append([x, y, float(s[9]), rate, 0])
if len(p_xy)>0:
p_xy = np.array(p_xy)
p_xy = p_xy[np.argsort(p_xy[:,2])]
p_xy = p_xy[np.where((p_xy[:,2]>nukeBright)&(p_xy[:,2]0:
for arg in ww[0]:
kb_xy[arg,-1] = p_xy[ii,2]-zpo
w_good = np.where(kb_xy[:,-1]!=0)
#load the stamps
with open(stamp_files[-1], 'rb') as han:
f = pickle.load(han)
if rots[counter%len(chips)]!=0:# and rots[counter%len(chips)]!=2:
f = np.rot90(f, k=-rots[counter%len(chips)], axes=(1, 2))
counter+=1
### clip to avoid the crazy min pixel values
f = np.clip(f, -3500., np.max(f))
f_med = np.nanmedian(f)
f = f[w_good]
kb_xy = kb_xy[w_good]
fs.append(f)
kb_xys.append(kb_xy)
zeropoints.append(np.zeros(len(f), dtype='float64')+zpo)
print(v, c,f.shape)
sns_frames = np.concatenate(fs)
kb_xys = np.concatenate(kb_xys)
zeropoints = np.concatenate(zeropoints)
print('Total sources:', zeropoints.shape)
2022-08-01-AS1_July 00 (45, 43, 43) 2022-08-01-AS1_July 01 (63, 43, 43) 2022-08-01-AS1_July 02 (55, 43, 43) 2022-08-01-AS1_July 03 (54, 43, 43) 2022-08-01-AS1_July 04 (58, 43, 43) 2022-08-01-AS1_July 05 (59, 43, 43) 2022-08-01-AS1_July 06 (70, 43, 43) 2022-08-01-AS1_July 07 (51, 43, 43) 2022-08-01-AS1_July 08 (64, 43, 43) 2022-08-01-AS1_July 09 (59, 43, 43) 2022-08-01-AS1_July 10 (58, 43, 43) 2022-08-01-AS1_July 11 (58, 43, 43) 2022-08-01-AS1_July 12 (60, 43, 43) 2022-08-01-AS1_July 13 (45, 43, 43) 2022-08-01-AS1_July 14 (65, 43, 43) 2022-08-01-AS1_July 15 (57, 43, 43) 2022-08-01-AS1_July 16 (63, 43, 43) 2022-08-01-AS1_July 17 (55, 43, 43) 2022-08-01-AS1_July 18 (62, 43, 43) 2022-08-01-AS1_July 19 (62, 43, 43) 2022-08-01-AS1_July 20 (58, 43, 43) 2022-08-01-AS1_July 21 (57, 43, 43) 2022-08-01-AS1_July 22 (51, 43, 43) 2022-08-01-AS1_July 23 (53, 43, 43) 2022-08-01-AS1_July 24 (49, 43, 43) 2022-08-01-AS1_July 25 (54, 43, 43) 2022-08-01-AS1_July 26 (50, 43, 43) 2022-08-01-AS1_July 27 (41, 43, 43) 2022-08-01-AS1_July 28 (48, 43, 43) 2022-08-01-AS1_July 29 (60, 43, 43) 2022-08-01-AS1_July 30 (51, 43, 43) 2022-08-01-AS1_July 31 (46, 43, 43) 2022-08-01-AS1_July 32 (50, 43, 43) 2022-08-01-AS1_July 33 (48, 43, 43) 2022-08-01-AS1_July 34 (58, 43, 43) 2022-08-01-AS1_July 35 (45, 43, 43) 2022-08-01-AS1_July 36 (56, 43, 43) 2022-08-01-AS1_July 37 (53, 43, 43) 2022-08-01-AS1_July 38 (53, 43, 43) 2022-08-01-AS1_July 39 (56, 43, 43) 2022-08-22-AS2 00 (57, 43, 43) 2022-08-22-AS2 01 (46, 43, 43) 2022-08-22-AS2 02 (54, 43, 43) 2022-08-22-AS2 03 (55, 43, 43) 2022-08-22-AS2 04 (63, 43, 43) 2022-08-22-AS2 05 (56, 43, 43) 2022-08-22-AS2 06 (72, 43, 43) 2022-08-22-AS2 07 (63, 43, 43) 2022-08-22-AS2 08 (45, 43, 43) 2022-08-22-AS2 09 (52, 43, 43) 2022-08-22-AS2 10 (53, 43, 43) 2022-08-22-AS2 11 (53, 43, 43) 2022-08-22-AS2 12 (57, 43, 43) 2022-08-22-AS2 13 (54, 43, 43) 2022-08-22-AS2 14 (69, 43, 43) 2022-08-22-AS2 15 (61, 43, 43) 2022-08-22-AS2 16 (46, 43, 43) 2022-08-22-AS2 17 (65, 43, 43) 2022-08-22-AS2 18 (59, 43, 43) 2022-08-22-AS2 19 (60, 43, 43) 2022-08-22-AS2 20 (54, 43, 43) 2022-08-22-AS2 21 (61, 43, 43) 2022-08-22-AS2 22 (59, 43, 43) 2022-08-22-AS2 23 (59, 43, 43) 2022-08-22-AS2 24 (51, 43, 43) 2022-08-22-AS2 25 (47, 43, 43) 2022-08-22-AS2 26 (57, 43, 43) 2022-08-22-AS2 27 (56, 43, 43) 2022-08-22-AS2 28 (55, 43, 43) 2022-08-22-AS2 29 (52, 43, 43) 2022-08-22-AS2 30 (65, 43, 43) 2022-08-22-AS2 31 (55, 43, 43) 2022-08-22-AS2 32 (43, 43, 43) 2022-08-22-AS2 33 (49, 43, 43) 2022-08-22-AS2 34 (55, 43, 43) 2022-08-22-AS2 35 (48, 43, 43) 2022-08-22-AS2 36 (60, 43, 43) 2022-08-22-AS2 37 (62, 43, 43) 2022-08-22-AS2 38 (60, 43, 43) 2022-08-22-AS2 39 (55, 43, 43) 2022-08-23-AS2 00 (62, 43, 43) 2022-08-23-AS2 01 (47, 43, 43) 2022-08-23-AS2 02 (57, 43, 43) 2022-08-23-AS2 03 (66, 43, 43) 2022-08-23-AS2 04 (60, 43, 43) 2022-08-23-AS2 05 (56, 43, 43) 2022-08-23-AS2 06 (71, 43, 43) 2022-08-23-AS2 07 (66, 43, 43) 2022-08-23-AS2 08 (46, 43, 43) 2022-08-23-AS2 09 (53, 43, 43) 2022-08-23-AS2 10 (52, 43, 43) 2022-08-23-AS2 11 (54, 43, 43) 2022-08-23-AS2 12 (62, 43, 43) 2022-08-23-AS2 13 (60, 43, 43) 2022-08-23-AS2 14 (72, 43, 43) 2022-08-23-AS2 15 (65, 43, 43) 2022-08-23-AS2 16 (58, 43, 43) 2022-08-23-AS2 17 (64, 43, 43) 2022-08-23-AS2 18 (57, 43, 43) 2022-08-23-AS2 19 (59, 43, 43) 2022-08-23-AS2 20 (54, 43, 43) 2022-08-23-AS2 21 (56, 43, 43) 2022-08-23-AS2 22 (55, 43, 43) 2022-08-23-AS2 23 (60, 43, 43) 2022-08-23-AS2 24 (52, 43, 43) 2022-08-23-AS2 25 (45, 43, 43) 2022-08-23-AS2 26 (52, 43, 43) 2022-08-23-AS2 27 (59, 43, 43) 2022-08-23-AS2 28 (46, 43, 43) 2022-08-23-AS2 29 (51, 43, 43) 2022-08-23-AS2 30 (64, 43, 43) 2022-08-23-AS2 31 (54, 43, 43) 2022-08-23-AS2 32 (45, 43, 43) 2022-08-23-AS2 33 (52, 43, 43) 2022-08-23-AS2 34 (54, 43, 43) 2022-08-23-AS2 35 (44, 43, 43) 2022-08-23-AS2 36 (58, 43, 43) 2022-08-23-AS2 37 (71, 43, 43) 2022-08-23-AS2 38 (57, 43, 43) 2022-08-23-AS2 39 (53, 43, 43) 2022-08-26-AS2 00 (60, 43, 43) 2022-08-26-AS2 01 (58, 43, 43) 2022-08-26-AS2 02 (61, 43, 43) 2022-08-26-AS2 03 (65, 43, 43) 2022-08-26-AS2 04 (58, 43, 43) 2022-08-26-AS2 05 (64, 43, 43) 2022-08-26-AS2 06 (80, 43, 43) 2022-08-26-AS2 07 (66, 43, 43) 2022-08-26-AS2 08 (58, 43, 43) 2022-08-26-AS2 09 (62, 43, 43) 2022-08-26-AS2 10 (52, 43, 43) 2022-08-26-AS2 11 (58, 43, 43) 2022-08-26-AS2 12 (64, 43, 43) 2022-08-26-AS2 13 (56, 43, 43) 2022-08-26-AS2 14 (64, 43, 43) 2022-08-26-AS2 15 (60, 43, 43) 2022-08-26-AS2 16 (66, 43, 43) 2022-08-26-AS2 17 (65, 43, 43) 2022-08-26-AS2 18 (70, 43, 43) 2022-08-26-AS2 19 (59, 43, 43) 2022-08-26-AS2 20 (71, 43, 43) 2022-08-26-AS2 21 (58, 43, 43) 2022-08-26-AS2 22 (58, 43, 43) 2022-08-26-AS2 23 (57, 43, 43) 2022-08-26-AS2 24 (50, 43, 43) 2022-08-26-AS2 25 (64, 43, 43) 2022-08-26-AS2 26 (54, 43, 43) 2022-08-26-AS2 27 (62, 43, 43) 2022-08-26-AS2 28 (61, 43, 43) 2022-08-26-AS2 29 (60, 43, 43) 2022-08-26-AS2 30 (62, 43, 43) 2022-08-26-AS2 31 (54, 43, 43) 2022-08-26-AS2 32 (50, 43, 43) 2022-08-26-AS2 33 (52, 43, 43) 2022-08-26-AS2 34 (58, 43, 43) 2022-08-26-AS2 35 (48, 43, 43) 2022-08-26-AS2 36 (59, 43, 43) 2022-08-26-AS2 37 (83, 43, 43) 2022-08-26-AS2 38 (68, 43, 43) 2022-08-26-AS2 39 (50, 43, 43) Total sources: (9137,)
In [4]:
Copied!
par_save_dic = {}
par_save_dic['visits'] = visits
par_save_dic['nukeBright'] = nukeBright
par_save_dic['nukeFaint'] = nukeFaint
par_save_dic['image_data_type'] = image_data_type
saved_par_dic = {
'visits': None,
'nukeBright': None,
'nukeFaint': None,
'image_data_type': None
}
par_save_dic = {}
par_save_dic['visits'] = visits
par_save_dic['nukeBright'] = nukeBright
par_save_dic['nukeFaint'] = nukeFaint
par_save_dic['image_data_type'] = image_data_type
saved_par_dic = {
'visits': None,
'nukeBright': None,
'nukeFaint': None,
'image_data_type': None
}
In [5]:
Copied!
###
#sort the labels and normalize
sns_labels = kb_xys[:, -1]
print('Min and max magnitudes:', np.min(sns_labels), np.max(sns_labels))
print()
sns_labels = 10.0**(-0.4*sns_labels)
Flux_std = np.std(sns_labels)
Flux_mean = np.mean(sns_labels)
print('Flux_mean/std:', Flux_mean, Flux_std)
sns_labels = sns_labels/Flux_std
print('Min and max normalized fluxes:', np.min(sns_labels), np.max(sns_labels))
###
#sort the labels and normalize
sns_labels = kb_xys[:, -1]
print('Min and max magnitudes:', np.min(sns_labels), np.max(sns_labels))
print()
sns_labels = 10.0**(-0.4*sns_labels)
Flux_std = np.std(sns_labels)
Flux_mean = np.mean(sns_labels)
print('Flux_mean/std:', Flux_mean, Flux_std)
sns_labels = sns_labels/Flux_std
print('Min and max normalized fluxes:', np.min(sns_labels), np.max(sns_labels))
Min and max magnitudes: -10.302000000000003 -5.863000000000003 Flux_mean/std: 3406.7642960502803 3380.2104471463504 Min and max normalized fluxes: 0.06550225634502951 3.9071155947150658
In [6]:
Copied!
#normalize the input image tensors
if useMedForNans:
med = np.nanmedian(sns_frames)
else:
med = 0.0
#mini = np.nanmin(sns_frames)
#sns_frames[np.where(np.isnan(sns_frames))] = mini
w_nan = np.where(np.isnan(sns_frames))
sns_frames[w_nan] = np.nan
normed_sns_frames = sns_frames
mean = np.nanmean(normed_sns_frames)
std = np.nanstd(normed_sns_frames)
normed_sns_frames -= mean
normed_sns_frames /= std
normed_sns_frames[w_nan] = 1.0 # 0.0
print('Normalized frame min and max:', np.nanmin(normed_sns_frames), np.nanmax(normed_sns_frames),'\n')
# expand the image data to shape (:, :, :, 1) for the CNN
normed_sns_frames = np.expand_dims(normed_sns_frames, axis=-1)
#normalize the input image tensors
if useMedForNans:
med = np.nanmedian(sns_frames)
else:
med = 0.0
#mini = np.nanmin(sns_frames)
#sns_frames[np.where(np.isnan(sns_frames))] = mini
w_nan = np.where(np.isnan(sns_frames))
sns_frames[w_nan] = np.nan
normed_sns_frames = sns_frames
mean = np.nanmean(normed_sns_frames)
std = np.nanstd(normed_sns_frames)
normed_sns_frames -= mean
normed_sns_frames /= std
normed_sns_frames[w_nan] = 1.0 # 0.0
print('Normalized frame min and max:', np.nanmin(normed_sns_frames), np.nanmax(normed_sns_frames),'\n')
# expand the image data to shape (:, :, :, 1) for the CNN
normed_sns_frames = np.expand_dims(normed_sns_frames, axis=-1)
Normalized frame min and max: -16.482555 591.32715
Augment the data.¶
Augmentations provide are shuffle: 1 pixel shifts in the cardinal directions; double_flip: flipping along the two axes; and rotate: rotating the images at 90, 180, and 270 degrees.
In [7]:
Copied!
if shuffle_augment:
a = np.copy(normed_sns_frames)
b = np.copy(normed_sns_frames)
c = np.copy(normed_sns_frames)
d = np.copy(normed_sns_frames)
a[:,:-1,:,:] = normed_sns_frames[:,1:,:,:]
b[:,1:,:,:] = normed_sns_frames[:,:-1,:,:]
c[:,:, :-1,:] = normed_sns_frames[:,:,1:,:]
d[:,:, 1:,:] = normed_sns_frames[:,:,:-1,:]
normed_sns_frames = np.concatenate([normed_sns_frames,a,b,c,d])#, A, B, C, D])
sns_labels = np.concatenate([sns_labels, np.copy(sns_labels), np.copy(sns_labels), np.copy(sns_labels), np.copy(sns_labels)])#, np.copy(sns_labels), np.copy(sns_labels), np.copy(sns_labels), np.copy(sns_labels)])
zeropoints = np.concatenate([zeropoints, np.copy(zeropoints), np.copy(zeropoints), np.copy(zeropoints), np.copy(zeropoints)])
kb_xys = np.concatenate([kb_xys, np.copy(kb_xys), np.copy(kb_xys), np.copy(kb_xys), np.copy(kb_xys)])
if double_flip:
normed_sns_frames = np.concatenate([normed_sns_frames, normed_sns_frames[:, ::-1, ::-1, :]])
sns_labels = np.concatenate([sns_labels, sns_labels])
zeropoints = np.concatenate([zeropoints, np.copy(zeropoints)])
kb_xys = np.concatenate([kb_xys, np.copy(kb_xys)])
if rotate_augment:
normed_sns_frames = np.concatenate([normed_sns_frames, np.rot90(normed_sns_frames, k=1, axes=(1,2)), np.rot90(normed_sns_frames, k=-1, axes=(1,2))])
sns_labels = np.concatenate([sns_labels, sns_labels, sns_labels])
zeropoints = np.concatenate([zeropoints, np.copy(zeropoints), np.copy(zeropoints)])
kb_xys = np.concatenate([kb_xys, np.copy(kb_xys), np.copy(kb_xys)])
if shuffle_augment:
a = np.copy(normed_sns_frames)
b = np.copy(normed_sns_frames)
c = np.copy(normed_sns_frames)
d = np.copy(normed_sns_frames)
a[:,:-1,:,:] = normed_sns_frames[:,1:,:,:]
b[:,1:,:,:] = normed_sns_frames[:,:-1,:,:]
c[:,:, :-1,:] = normed_sns_frames[:,:,1:,:]
d[:,:, 1:,:] = normed_sns_frames[:,:,:-1,:]
normed_sns_frames = np.concatenate([normed_sns_frames,a,b,c,d])#, A, B, C, D])
sns_labels = np.concatenate([sns_labels, np.copy(sns_labels), np.copy(sns_labels), np.copy(sns_labels), np.copy(sns_labels)])#, np.copy(sns_labels), np.copy(sns_labels), np.copy(sns_labels), np.copy(sns_labels)])
zeropoints = np.concatenate([zeropoints, np.copy(zeropoints), np.copy(zeropoints), np.copy(zeropoints), np.copy(zeropoints)])
kb_xys = np.concatenate([kb_xys, np.copy(kb_xys), np.copy(kb_xys), np.copy(kb_xys), np.copy(kb_xys)])
if double_flip:
normed_sns_frames = np.concatenate([normed_sns_frames, normed_sns_frames[:, ::-1, ::-1, :]])
sns_labels = np.concatenate([sns_labels, sns_labels])
zeropoints = np.concatenate([zeropoints, np.copy(zeropoints)])
kb_xys = np.concatenate([kb_xys, np.copy(kb_xys)])
if rotate_augment:
normed_sns_frames = np.concatenate([normed_sns_frames, np.rot90(normed_sns_frames, k=1, axes=(1,2)), np.rot90(normed_sns_frames, k=-1, axes=(1,2))])
sns_labels = np.concatenate([sns_labels, sns_labels, sns_labels])
zeropoints = np.concatenate([zeropoints, np.copy(zeropoints), np.copy(zeropoints)])
kb_xys = np.concatenate([kb_xys, np.copy(kb_xys), np.copy(kb_xys)])
In [8]:
Copied!
## it's always a good idea to check for nans
print(np.where(np.isnan(normed_sns_frames)))
print(np.where(np.isnan(sns_labels)))
print(np.where(np.isnan(zeropoints)))
## it's always a good idea to check for nans
print(np.where(np.isnan(normed_sns_frames)))
print(np.where(np.isnan(sns_labels)))
print(np.where(np.isnan(zeropoints)))
(array([], dtype=int64), array([], dtype=int64), array([], dtype=int64), array([], dtype=int64)) (array([], dtype=int64),) (array([], dtype=int64),)
Training and Test/Validation Samples¶
In [9]:
Copied!
## always a good idea to confirm that the training x and y arrays have the same # of elements
print(sns_labels.shape, normed_sns_frames.shape)
test_index = (np.random.rand(int(len(sns_labels)*(test_fraction)))*len(sns_labels)).astype('int')
train_index = []
for i in range(len(sns_labels)):
if i not in test_index:
train_index.append(i)
train_index = np.array(train_index)
X_train, X_test = normed_sns_frames[train_index], normed_sns_frames[test_index]
y_train, y_test = sns_labels[train_index], sns_labels[test_index]
Z_train, Z_test = zeropoints[train_index], zeropoints[test_index]
#del normed_sns_frames
print(X_train.shape)
print('Number of images in the training sample: ', X_train.shape[0])
print('Number of images in the test sample: ', X_test.shape[0])
## always a good idea to confirm that the training x and y arrays have the same # of elements
print(sns_labels.shape, normed_sns_frames.shape)
test_index = (np.random.rand(int(len(sns_labels)*(test_fraction)))*len(sns_labels)).astype('int')
train_index = []
for i in range(len(sns_labels)):
if i not in test_index:
train_index.append(i)
train_index = np.array(train_index)
X_train, X_test = normed_sns_frames[train_index], normed_sns_frames[test_index]
y_train, y_test = sns_labels[train_index], sns_labels[test_index]
Z_train, Z_test = zeropoints[train_index], zeropoints[test_index]
#del normed_sns_frames
print(X_train.shape)
print('Number of images in the training sample: ', X_train.shape[0])
print('Number of images in the test sample: ', X_test.shape[0])
(274110,) (274110, 43, 43, 1) (260720, 43, 43, 1) Number of images in the training sample: 260720 Number of images in the test sample: 13705
Histogram of the input magnitudes.¶
Generate a set of sample weights. Look at the flux distribution, and then interpolate, and invert to generate weights. nbins=100 seems to result in a smooth flux_pred/flux_train and test curve. Extend the weights histogram to the min and max of sns_labels so to avoid interpolation issues.
In [10]:
Copied!
nbins = 25
fig = pyl.figure(figsize=(10,10))
sp = fig.add_subplot(111)
(y, x) = pyl.hist(sns_labels, bins=nbins)[:2]
pyl.plot([0.35, 0.35], sp.get_ylim())
bins = (x[1:]+x[:-1])/2.0
pyl.scatter(bins, y, c='r',zorder=10)
pyl.show()
hist_weights = np.max(y)/(y*nbins)
hist_weights = np.concatenate([np.array([hist_weights[0]]),
hist_weights,
np.array([hist_weights[-1]])])
bins = np.concatenate([np.array([np.min(sns_labels)]),
bins,
np.array([np.max(sns_labels)])])
hist_weights[np.where(np.isinf(hist_weights))] = 0.0
fig = pyl.figure(figsize=(10,10))
pyl.scatter(bins, hist_weights)
pyl.show()
weight_func = interp.interp1d(bins, hist_weights)
sample_weights = weight_func(sns_labels).astype('float32')
mean_sample_weights = np.mean(sample_weights)
sample_weights /= mean_sample_weights
train_weights = sample_weights[train_index]
test_weights = sample_weights[test_index]
nbins = 25
fig = pyl.figure(figsize=(10,10))
sp = fig.add_subplot(111)
(y, x) = pyl.hist(sns_labels, bins=nbins)[:2]
pyl.plot([0.35, 0.35], sp.get_ylim())
bins = (x[1:]+x[:-1])/2.0
pyl.scatter(bins, y, c='r',zorder=10)
pyl.show()
hist_weights = np.max(y)/(y*nbins)
hist_weights = np.concatenate([np.array([hist_weights[0]]),
hist_weights,
np.array([hist_weights[-1]])])
bins = np.concatenate([np.array([np.min(sns_labels)]),
bins,
np.array([np.max(sns_labels)])])
hist_weights[np.where(np.isinf(hist_weights))] = 0.0
fig = pyl.figure(figsize=(10,10))
pyl.scatter(bins, hist_weights)
pyl.show()
weight_func = interp.interp1d(bins, hist_weights)
sample_weights = weight_func(sns_labels).astype('float32')
mean_sample_weights = np.mean(sample_weights)
sample_weights /= mean_sample_weights
train_weights = sample_weights[train_index]
test_weights = sample_weights[test_index]
Fitting the Neural Network Model¶
Here we train the neural network model defined above. The training session will output loss, and accuracy at each epoch. We also plot the progression of the loss and accuracy with respect to training epochs after the training is completed.
In [11]:
Copied!
backend.clear_session()
gc.collect()
cn_model = convnet_model(X_train.shape[1:],
learning_rate = learning_rate*4,
num_dense_nodes = num_dense_nodes,
num_dense_layers = num_dense_layers,
num_models = num_models, num_filters = num_filters)
cn_model.compile()
cn_model.model_.summary()
start = time.time()
cn_model.train_models(X_train, y_train, sample_weights = train_weights,
train_epochs = train_epochs,
batch_size=int(batch_size),
useSampleWeights = useSampleWeights)
end = time.time()
print('Process completed in', round(end-start, 2), ' seconds')
backend.clear_session()
gc.collect()
cn_model = convnet_model(X_train.shape[1:],
learning_rate = learning_rate*4,
num_dense_nodes = num_dense_nodes,
num_dense_layers = num_dense_layers,
num_models = num_models, num_filters = num_filters)
cn_model.compile()
cn_model.model_.summary()
start = time.time()
cn_model.train_models(X_train, y_train, sample_weights = train_weights,
train_epochs = train_epochs,
batch_size=int(batch_size),
useSampleWeights = useSampleWeights)
end = time.time()
print('Process completed in', round(end-start, 2), ' seconds')
2024-04-18 18:23:05.361370: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero 2024-04-18 18:23:05.374168: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero 2024-04-18 18:23:05.376408: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero 2024-04-18 18:23:05.379878: I tensorflow/core/platform/cpu_feature_guard.cc:142] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations: SSE4.1 SSE4.2 AVX AVX2 AVX512F FMA To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags. 2024-04-18 18:23:05.403570: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero 2024-04-18 18:23:05.405917: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero 2024-04-18 18:23:05.408033: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero 2024-04-18 18:23:07.365550: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero 2024-04-18 18:23:07.367215: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero 2024-04-18 18:23:07.368635: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero 2024-04-18 18:23:07.369911: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1510] Created device /job:localhost/replica:0/task:0/device:GPU:0 with 18043 MB memory: -> device: 0, name: NVIDIA A100-PCIE-40GB MIG 3g.20gb, pci bus id: 0000:00:07.0, compute capability: 8.0
Model: "model_4"
__________________________________________________________________________________________________
Layer (type) Output Shape Param # Connected to
==================================================================================================
input (InputLayer) [(None, 43, 43, 1)] 0
__________________________________________________________________________________________________
conv2d_12 (Conv2D) (None, 41, 41, 6) 60 input[0][0]
__________________________________________________________________________________________________
max_pooling2d_12 (MaxPooling2D) (None, 21, 21, 6) 0 conv2d_12[0][0]
__________________________________________________________________________________________________
conv2d_13 (Conv2D) (None, 19, 19, 6) 330 max_pooling2d_12[0][0]
__________________________________________________________________________________________________
max_pooling2d_13 (MaxPooling2D) (None, 9, 9, 6) 0 conv2d_13[0][0]
__________________________________________________________________________________________________
conv2d_14 (Conv2D) (None, 7, 7, 3) 165 max_pooling2d_13[0][0]
__________________________________________________________________________________________________
max_pooling2d_14 (MaxPooling2D) (None, 3, 3, 3) 0 conv2d_14[0][0]
__________________________________________________________________________________________________
flatten_4 (Flatten) (None, 27) 0 max_pooling2d_14[0][0]
__________________________________________________________________________________________________
dense_8 (Dense) (None, 8) 224 flatten_4[0][0]
__________________________________________________________________________________________________
dense_9 (Dense) (None, 8) 72 dense_8[0][0]
__________________________________________________________________________________________________
mu (Dense) (None, 1) 9 dense_9[0][0]
__________________________________________________________________________________________________
sigma (Dense) (None, 1) 9 dense_9[0][0]
__________________________________________________________________________________________________
outputs (Concatenate) (None, 2) 0 mu[0][0]
sigma[0][0]
==================================================================================================
Total params: 869
Trainable params: 869
Non-trainable params: 0
__________________________________________________________________________________________________
Training model 1 of 5.
2024-04-18 18:23:44.519403: I tensorflow/compiler/mlir/mlir_graph_optimization_pass.cc:185] None of the MLIR Optimization Passes are enabled (registered 2)
Epoch 1/250
2024-04-18 18:23:57.792846: I tensorflow/stream_executor/cuda/cuda_dnn.cc:369] Loaded cuDNN version 8201 2024-04-18 18:24:00.068096: I tensorflow/core/platform/default/subprocess.cc:304] Start cannot spawn child process: No such file or directory 2024-04-18 18:24:01.711546: I tensorflow/stream_executor/cuda/cuda_blas.cc:1760] TensorFloat-32 will be used for the matrix multiplication. This will only be logged once.
64/64 [==============================] - 22s 44ms/step - loss: 0.9640 - mean_absolute_error: 0.6547 Epoch 2/250 64/64 [==============================] - 2s 37ms/step - loss: -0.6215 - mean_absolute_error: 0.4630 Epoch 3/250 64/64 [==============================] - 2s 36ms/step - loss: -1.2326 - mean_absolute_error: 0.4628 Epoch 4/250 64/64 [==============================] - 2s 34ms/step - loss: -1.5025 - mean_absolute_error: 0.4792 Epoch 5/250 64/64 [==============================] - 2s 34ms/step - loss: -1.6371 - mean_absolute_error: 0.4847 Epoch 6/250 64/64 [==============================] - 2s 34ms/step - loss: -1.7096 - mean_absolute_error: 0.4877 Epoch 7/250 64/64 [==============================] - 2s 34ms/step - loss: -1.7425 - mean_absolute_error: 0.4895 Epoch 8/250 64/64 [==============================] - 2s 34ms/step - loss: -1.7844 - mean_absolute_error: 0.4900 Epoch 9/250 64/64 [==============================] - 2s 34ms/step - loss: -1.8057 - mean_absolute_error: 0.4918 Epoch 10/250 64/64 [==============================] - 2s 35ms/step - loss: -1.8291 - mean_absolute_error: 0.4919 Epoch 11/250 64/64 [==============================] - 2s 34ms/step - loss: -1.8479 - mean_absolute_error: 0.4921 Epoch 12/250 64/64 [==============================] - 2s 36ms/step - loss: -1.8616 - mean_absolute_error: 0.4930 Epoch 13/250 64/64 [==============================] - 2s 37ms/step - loss: -1.8891 - mean_absolute_error: 0.4938 Epoch 14/250 64/64 [==============================] - 2s 35ms/step - loss: -1.9022 - mean_absolute_error: 0.4937 Epoch 15/250 64/64 [==============================] - 2s 34ms/step - loss: -1.9188 - mean_absolute_error: 0.4939 Epoch 16/250 64/64 [==============================] - 2s 34ms/step - loss: -1.9297 - mean_absolute_error: 0.4941 Epoch 17/250 64/64 [==============================] - 2s 34ms/step - loss: -1.9376 - mean_absolute_error: 0.4943 Epoch 18/250 64/64 [==============================] - 2s 34ms/step - loss: -1.9468 - mean_absolute_error: 0.4944 Epoch 19/250 64/64 [==============================] - 2s 34ms/step - loss: -1.9560 - mean_absolute_error: 0.4947 Epoch 20/250 64/64 [==============================] - 2s 34ms/step - loss: -1.9631 - mean_absolute_error: 0.4950 Epoch 21/250 64/64 [==============================] - 2s 34ms/step - loss: -1.9717 - mean_absolute_error: 0.4951 Epoch 22/250 64/64 [==============================] - 2s 34ms/step - loss: -1.9819 - mean_absolute_error: 0.4952 Epoch 23/250 64/64 [==============================] - 2s 35ms/step - loss: -1.9829 - mean_absolute_error: 0.4955 Epoch 24/250 64/64 [==============================] - 2s 33ms/step - loss: -1.9926 - mean_absolute_error: 0.4959 Epoch 25/250 64/64 [==============================] - 2s 35ms/step - loss: -2.0003 - mean_absolute_error: 0.4961 Epoch 26/250 64/64 [==============================] - 2s 36ms/step - loss: -2.0074 - mean_absolute_error: 0.4963 Epoch 27/250 64/64 [==============================] - 2s 33ms/step - loss: -2.0116 - mean_absolute_error: 0.4964 Epoch 28/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0166 - mean_absolute_error: 0.4965 Epoch 29/250 64/64 [==============================] - 2s 33ms/step - loss: -2.0263 - mean_absolute_error: 0.4970 Epoch 30/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0285 - mean_absolute_error: 0.4967 Epoch 31/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0370 - mean_absolute_error: 0.4972 Epoch 32/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0359 - mean_absolute_error: 0.4972 Epoch 33/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0418 - mean_absolute_error: 0.4971 Epoch 34/250 64/64 [==============================] - 2s 33ms/step - loss: -2.0490 - mean_absolute_error: 0.4977 Epoch 35/250 64/64 [==============================] - 2s 35ms/step - loss: -2.0527 - mean_absolute_error: 0.4977 Epoch 36/250 64/64 [==============================] - 2s 35ms/step - loss: -2.0556 - mean_absolute_error: 0.4979 Epoch 37/250 64/64 [==============================] - 2s 36ms/step - loss: -2.0546 - mean_absolute_error: 0.4980 Epoch 38/250 64/64 [==============================] - 2s 35ms/step - loss: -2.0641 - mean_absolute_error: 0.4981 Epoch 39/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0648 - mean_absolute_error: 0.4982 Epoch 40/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0735 - mean_absolute_error: 0.4984 Epoch 41/250 64/64 [==============================] - 2s 33ms/step - loss: -2.0736 - mean_absolute_error: 0.4984 Epoch 42/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0720 - mean_absolute_error: 0.4987 Epoch 43/250 64/64 [==============================] - 2s 32ms/step - loss: -2.0822 - mean_absolute_error: 0.4989 Epoch 44/250 64/64 [==============================] - 2s 33ms/step - loss: -2.0840 - mean_absolute_error: 0.4988 Epoch 45/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0878 - mean_absolute_error: 0.4989 Epoch 46/250 64/64 [==============================] - 2s 33ms/step - loss: -2.0858 - mean_absolute_error: 0.4990 1s - l Epoch 47/250 64/64 [==============================] - 2s 33ms/step - loss: -2.0921 - mean_absolute_error: 0.4989 Epoch 48/250 64/64 [==============================] - 2s 33ms/step - loss: -2.0904 - mean_absolute_error: 0.4990 Epoch 49/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0953 - mean_absolute_error: 0.4991 Epoch 50/250 64/64 [==============================] - 2s 35ms/step - loss: -2.0992 - mean_absolute_error: 0.4993 Epoch 51/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1031 - mean_absolute_error: 0.4992 Epoch 52/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1032 - mean_absolute_error: 0.4995 Epoch 53/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1045 - mean_absolute_error: 0.4994 Epoch 54/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1052 - mean_absolute_error: 0.4994 Epoch 55/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1041 - mean_absolute_error: 0.4994 Epoch 56/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1101 - mean_absolute_error: 0.4994 Epoch 57/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1151 - mean_absolute_error: 0.4996 Epoch 58/250 64/64 [==============================] - 2s 32ms/step - loss: -2.1137 - mean_absolute_error: 0.4997 Epoch 59/250 64/64 [==============================] - 2s 32ms/step - loss: -2.1200 - mean_absolute_error: 0.4996 Epoch 60/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0579 - mean_absolute_error: 0.4998 Epoch 61/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0566 - mean_absolute_error: 0.4993 Epoch 62/250 64/64 [==============================] - 2s 35ms/step - loss: -2.0612 - mean_absolute_error: 0.4992 Epoch 63/250 64/64 [==============================] - 2s 35ms/step - loss: -2.0595 - mean_absolute_error: 0.4991 Epoch 64/250 64/64 [==============================] - 2s 35ms/step - loss: -2.0524 - mean_absolute_error: 0.4990 Epoch 65/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0557 - mean_absolute_error: 0.4986 Epoch 66/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0605 - mean_absolute_error: 0.4989 Epoch 67/250 64/64 [==============================] - 2s 35ms/step - loss: -2.0536 - mean_absolute_error: 0.4984 Epoch 68/250 64/64 [==============================] - 2s 37ms/step - loss: -2.0549 - mean_absolute_error: 0.4985 Epoch 69/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0585 - mean_absolute_error: 0.4984 Epoch 70/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0578 - mean_absolute_error: 0.4987 Epoch 71/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0572 - mean_absolute_error: 0.4979 Epoch 72/250 64/64 [==============================] - 2s 35ms/step - loss: -2.0563 - mean_absolute_error: 0.4980 Epoch 73/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0520 - mean_absolute_error: 0.4980 Epoch 74/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0619 - mean_absolute_error: 0.4980 Epoch 75/250 64/64 [==============================] - 2s 33ms/step - loss: -2.0577 - mean_absolute_error: 0.4982 Epoch 76/250 64/64 [==============================] - 2s 35ms/step - loss: -2.0500 - mean_absolute_error: 0.4979 Epoch 77/250 64/64 [==============================] - 2s 33ms/step - loss: -2.0502 - mean_absolute_error: 0.4978 Epoch 78/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0452 - mean_absolute_error: 0.4983 Epoch 79/250 64/64 [==============================] - 2s 39ms/step - loss: -2.0384 - mean_absolute_error: 0.4991 Epoch 80/250 64/64 [==============================] - 2s 36ms/step - loss: -2.1244 - mean_absolute_error: 0.4981 Epoch 81/250 64/64 [==============================] - 2s 37ms/step - loss: -2.1309 - mean_absolute_error: 0.4982 Epoch 82/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1344 - mean_absolute_error: 0.4986 Epoch 83/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1280 - mean_absolute_error: 0.4987 Epoch 84/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1312 - mean_absolute_error: 0.4989 Epoch 85/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1403 - mean_absolute_error: 0.4989 Epoch 86/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1412 - mean_absolute_error: 0.4991 Epoch 87/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1380 - mean_absolute_error: 0.4990 Epoch 88/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1361 - mean_absolute_error: 0.4992 Epoch 89/250 64/64 [==============================] - 2s 36ms/step - loss: -2.1329 - mean_absolute_error: 0.4990 Epoch 90/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1337 - mean_absolute_error: 0.4991 Epoch 91/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1407 - mean_absolute_error: 0.4993 Epoch 92/250 64/64 [==============================] - 2s 36ms/step - loss: -2.1437 - mean_absolute_error: 0.4990 Epoch 93/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1472 - mean_absolute_error: 0.4996 Epoch 94/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1472 - mean_absolute_error: 0.4994 Epoch 95/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1474 - mean_absolute_error: 0.4999 Epoch 96/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1479 - mean_absolute_error: 0.4994 Epoch 97/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1515 - mean_absolute_error: 0.4996 Epoch 98/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1565 - mean_absolute_error: 0.4995 Epoch 99/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1616 - mean_absolute_error: 0.4996 Epoch 100/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1609 - mean_absolute_error: 0.4999 Epoch 101/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1647 - mean_absolute_error: 0.4996 Epoch 102/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1641 - mean_absolute_error: 0.4998 Epoch 103/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1584 - mean_absolute_error: 0.4999 Epoch 104/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1595 - mean_absolute_error: 0.5000 Epoch 105/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1583 - mean_absolute_error: 0.4996 Epoch 106/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1659 - mean_absolute_error: 0.5001 Epoch 107/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1675 - mean_absolute_error: 0.5003 Epoch 108/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1676 - mean_absolute_error: 0.5003 Epoch 109/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1720 - mean_absolute_error: 0.5002 Epoch 110/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1628 - mean_absolute_error: 0.5004 Epoch 111/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1547 - mean_absolute_error: 0.5004 Epoch 112/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1557 - mean_absolute_error: 0.5002 Epoch 113/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1593 - mean_absolute_error: 0.5006 Epoch 114/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1656 - mean_absolute_error: 0.5005 Epoch 115/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1649 - mean_absolute_error: 0.5006 Epoch 116/250 64/64 [==============================] - 2s 36ms/step - loss: -2.1679 - mean_absolute_error: 0.5005 Epoch 117/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1666 - mean_absolute_error: 0.5005 Epoch 118/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1615 - mean_absolute_error: 0.5008 Epoch 119/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1570 - mean_absolute_error: 0.5005 Epoch 120/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1591 - mean_absolute_error: 0.5008 Epoch 121/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1577 - mean_absolute_error: 0.5006 Epoch 122/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1686 - mean_absolute_error: 0.5011 Epoch 123/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1777 - mean_absolute_error: 0.5011 Epoch 124/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1807 - mean_absolute_error: 0.5010 Epoch 125/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1818 - mean_absolute_error: 0.5008 Epoch 126/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1818 - mean_absolute_error: 0.5009 Epoch 127/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1745 - mean_absolute_error: 0.5012 Epoch 128/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1728 - mean_absolute_error: 0.5012 Epoch 129/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1727 - mean_absolute_error: 0.5012 Epoch 130/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1750 - mean_absolute_error: 0.5008 Epoch 131/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1785 - mean_absolute_error: 0.5010 Epoch 132/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1830 - mean_absolute_error: 0.5012 Epoch 133/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1872 - mean_absolute_error: 0.5010 Epoch 134/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1856 - mean_absolute_error: 0.5011 Epoch 135/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1898 - mean_absolute_error: 0.5010 Epoch 136/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1435 - mean_absolute_error: 0.5016 Epoch 137/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0827 - mean_absolute_error: 0.5025 Epoch 138/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0894 - mean_absolute_error: 0.5025 Epoch 139/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0799 - mean_absolute_error: 0.5030 Epoch 140/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0969 - mean_absolute_error: 0.5019 Epoch 141/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0927 - mean_absolute_error: 0.5022 Epoch 142/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1880 - mean_absolute_error: 0.5031 Epoch 143/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1596 - mean_absolute_error: 0.5024 Epoch 144/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1530 - mean_absolute_error: 0.5026 Epoch 145/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1659 - mean_absolute_error: 0.5024 Epoch 146/250 64/64 [==============================] - 2s 32ms/step - loss: -2.1629 - mean_absolute_error: 0.5026 Epoch 147/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1538 - mean_absolute_error: 0.5032 Epoch 148/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1576 - mean_absolute_error: 0.5026 Epoch 149/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1672 - mean_absolute_error: 0.5026 Epoch 150/250 64/64 [==============================] - 2s 36ms/step - loss: -2.1681 - mean_absolute_error: 0.5027 Epoch 151/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1668 - mean_absolute_error: 0.5026 Epoch 152/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1693 - mean_absolute_error: 0.5027 Epoch 153/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1713 - mean_absolute_error: 0.5028 Epoch 154/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1753 - mean_absolute_error: 0.5029 Epoch 155/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1799 - mean_absolute_error: 0.5031 Epoch 156/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1809 - mean_absolute_error: 0.5030 Epoch 157/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1843 - mean_absolute_error: 0.5030 Epoch 158/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1861 - mean_absolute_error: 0.5031 Epoch 159/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1864 - mean_absolute_error: 0.5031 Epoch 160/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1870 - mean_absolute_error: 0.5030 Epoch 161/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1867 - mean_absolute_error: 0.5031 Epoch 162/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1904 - mean_absolute_error: 0.5033 Epoch 163/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1878 - mean_absolute_error: 0.5033 Epoch 164/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1901 - mean_absolute_error: 0.5031 Epoch 165/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1940 - mean_absolute_error: 0.5031 Epoch 166/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1932 - mean_absolute_error: 0.5031 Epoch 167/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1927 - mean_absolute_error: 0.5030 Epoch 168/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1948 - mean_absolute_error: 0.5030 Epoch 169/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1934 - mean_absolute_error: 0.5031 Epoch 170/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1928 - mean_absolute_error: 0.5031 Epoch 171/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1970 - mean_absolute_error: 0.5030 Epoch 172/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1954 - mean_absolute_error: 0.5029 Epoch 173/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1968 - mean_absolute_error: 0.5031 0s - loss: -2.1947 - mean_absolute_err Epoch 174/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1960 - mean_absolute_error: 0.5030 Epoch 175/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1973 - mean_absolute_error: 0.5032 Epoch 176/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1967 - mean_absolute_error: 0.5029 Epoch 177/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1980 - mean_absolute_error: 0.5029 Epoch 178/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2003 - mean_absolute_error: 0.5028 Epoch 179/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1969 - mean_absolute_error: 0.5029 Epoch 180/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1941 - mean_absolute_error: 0.5029 Epoch 181/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1894 - mean_absolute_error: 0.5030 Epoch 182/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1840 - mean_absolute_error: 0.5033 Epoch 183/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1915 - mean_absolute_error: 0.5028 Epoch 184/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1851 - mean_absolute_error: 0.5029 Epoch 185/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1897 - mean_absolute_error: 0.5038 Epoch 186/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1918 - mean_absolute_error: 0.5036 Epoch 187/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1923 - mean_absolute_error: 0.5033 Epoch 188/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1867 - mean_absolute_error: 0.5033 Epoch 189/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1890 - mean_absolute_error: 0.5030 Epoch 190/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1758 - mean_absolute_error: 0.5029 Epoch 191/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1756 - mean_absolute_error: 0.5026 Epoch 192/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1792 - mean_absolute_error: 0.5028 Epoch 193/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1803 - mean_absolute_error: 0.5027 Epoch 194/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1798 - mean_absolute_error: 0.5025 Epoch 195/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1808 - mean_absolute_error: 0.5024 Epoch 196/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1784 - mean_absolute_error: 0.5027 Epoch 197/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1714 - mean_absolute_error: 0.5024 Epoch 198/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1766 - mean_absolute_error: 0.5024 Epoch 199/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1707 - mean_absolute_error: 0.5026 Epoch 200/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1714 - mean_absolute_error: 0.5025 Epoch 201/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1728 - mean_absolute_error: 0.5024 Epoch 202/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1706 - mean_absolute_error: 0.5024 Epoch 203/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1737 - mean_absolute_error: 0.5023 Epoch 204/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1786 - mean_absolute_error: 0.5024 Epoch 205/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1730 - mean_absolute_error: 0.5021 Epoch 206/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1685 - mean_absolute_error: 0.5029 Epoch 207/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1733 - mean_absolute_error: 0.5030 Epoch 208/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1772 - mean_absolute_error: 0.5027 Epoch 209/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1748 - mean_absolute_error: 0.5023 Epoch 210/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1778 - mean_absolute_error: 0.5025 Epoch 211/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1778 - mean_absolute_error: 0.5022 Epoch 212/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1782 - mean_absolute_error: 0.5023 Epoch 213/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1800 - mean_absolute_error: 0.5023 Epoch 214/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1760 - mean_absolute_error: 0.5023 Epoch 215/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1803 - mean_absolute_error: 0.5024 Epoch 216/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1762 - mean_absolute_error: 0.5027 Epoch 217/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1787 - mean_absolute_error: 0.5020 Epoch 218/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1786 - mean_absolute_error: 0.5022 Epoch 219/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1840 - mean_absolute_error: 0.5023 Epoch 220/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1832 - mean_absolute_error: 0.5022 Epoch 221/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1835 - mean_absolute_error: 0.5023 Epoch 222/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1809 - mean_absolute_error: 0.5025 Epoch 223/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1849 - mean_absolute_error: 0.5026 Epoch 224/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1851 - mean_absolute_error: 0.5024 Epoch 225/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1865 - mean_absolute_error: 0.5023 Epoch 226/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1887 - mean_absolute_error: 0.5023 Epoch 227/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1854 - mean_absolute_error: 0.5024 Epoch 228/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1888 - mean_absolute_error: 0.5024 Epoch 229/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1882 - mean_absolute_error: 0.5022 Epoch 230/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1907 - mean_absolute_error: 0.5024 Epoch 231/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1889 - mean_absolute_error: 0.5023 Epoch 232/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1937 - mean_absolute_error: 0.5025 Epoch 233/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1904 - mean_absolute_error: 0.5024 Epoch 234/250 64/64 [==============================] - 2s 32ms/step - loss: -2.1912 - mean_absolute_error: 0.5026 Epoch 235/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1891 - mean_absolute_error: 0.5025 Epoch 236/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1869 - mean_absolute_error: 0.5027 Epoch 237/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1870 - mean_absolute_error: 0.5032 Epoch 238/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1885 - mean_absolute_error: 0.5033 Epoch 239/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1929 - mean_absolute_error: 0.5030 Epoch 240/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1969 - mean_absolute_error: 0.5032 Epoch 241/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1994 - mean_absolute_error: 0.5029 Epoch 242/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1950 - mean_absolute_error: 0.5033 Epoch 243/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1986 - mean_absolute_error: 0.5030 Epoch 244/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1984 - mean_absolute_error: 0.5029 Epoch 245/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1960 - mean_absolute_error: 0.5031 Epoch 246/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1953 - mean_absolute_error: 0.5032 Epoch 247/250 64/64 [==============================] - 2s 32ms/step - loss: -2.2026 - mean_absolute_error: 0.5032 Epoch 248/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1991 - mean_absolute_error: 0.5033 Epoch 249/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1984 - mean_absolute_error: 0.5035 Epoch 250/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2005 - mean_absolute_error: 0.5036 Training model 2 of 5. Epoch 1/250 64/64 [==============================] - 5s 47ms/step - loss: 0.1349 - mean_absolute_error: 0.5666 Epoch 2/250 64/64 [==============================] - 2s 33ms/step - loss: -1.4145 - mean_absolute_error: 0.4784 Epoch 3/250 64/64 [==============================] - 2s 33ms/step - loss: -1.8075 - mean_absolute_error: 0.4859 Epoch 4/250 64/64 [==============================] - 2s 34ms/step - loss: -1.8844 - mean_absolute_error: 0.4880 Epoch 5/250 64/64 [==============================] - 2s 33ms/step - loss: -1.9321 - mean_absolute_error: 0.4893 Epoch 6/250 64/64 [==============================] - 2s 33ms/step - loss: -1.9658 - mean_absolute_error: 0.4901 Epoch 7/250 64/64 [==============================] - 2s 33ms/step - loss: -1.9931 - mean_absolute_error: 0.4910 Epoch 8/250 64/64 [==============================] - 2s 35ms/step - loss: -2.0142 - mean_absolute_error: 0.4917 Epoch 9/250 64/64 [==============================] - 2s 33ms/step - loss: -2.0343 - mean_absolute_error: 0.4926 Epoch 10/250 64/64 [==============================] - 2s 35ms/step - loss: -2.0551 - mean_absolute_error: 0.4933 Epoch 11/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0706 - mean_absolute_error: 0.4934 Epoch 12/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0936 - mean_absolute_error: 0.4943 Epoch 13/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1062 - mean_absolute_error: 0.4945 Epoch 14/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1159 - mean_absolute_error: 0.4948 Epoch 15/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1229 - mean_absolute_error: 0.4951 Epoch 16/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1313 - mean_absolute_error: 0.4953 Epoch 17/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1511 - mean_absolute_error: 0.4955 Epoch 18/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1621 - mean_absolute_error: 0.4959 Epoch 19/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1673 - mean_absolute_error: 0.4964 Epoch 20/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1821 - mean_absolute_error: 0.4964 Epoch 21/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1893 - mean_absolute_error: 0.4970 Epoch 22/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1722 - mean_absolute_error: 0.4967 Epoch 23/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1792 - mean_absolute_error: 0.4968 Epoch 24/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1826 - mean_absolute_error: 0.4969 Epoch 25/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1884 - mean_absolute_error: 0.4971 Epoch 26/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2143 - mean_absolute_error: 0.4974 Epoch 27/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2167 - mean_absolute_error: 0.4976 Epoch 28/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2245 - mean_absolute_error: 0.4978 Epoch 29/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2304 - mean_absolute_error: 0.4979 Epoch 30/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2374 - mean_absolute_error: 0.4980 Epoch 31/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2378 - mean_absolute_error: 0.4981 Epoch 32/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2402 - mean_absolute_error: 0.4980 Epoch 33/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2440 - mean_absolute_error: 0.4983 Epoch 34/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2351 - mean_absolute_error: 0.4981 Epoch 35/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2289 - mean_absolute_error: 0.4982 Epoch 36/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2322 - mean_absolute_error: 0.4982 Epoch 37/250 64/64 [==============================] - 2s 37ms/step - loss: -2.2322 - mean_absolute_error: 0.4984 Epoch 38/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2354 - mean_absolute_error: 0.4983 Epoch 39/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2373 - mean_absolute_error: 0.4982 Epoch 40/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2396 - mean_absolute_error: 0.4982 Epoch 41/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2408 - mean_absolute_error: 0.4983 Epoch 42/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2429 - mean_absolute_error: 0.4983 Epoch 43/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2433 - mean_absolute_error: 0.4983 Epoch 44/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2458 - mean_absolute_error: 0.4983 Epoch 45/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2473 - mean_absolute_error: 0.4983 Epoch 46/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2482 - mean_absolute_error: 0.4984 Epoch 47/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2500 - mean_absolute_error: 0.4984 Epoch 48/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2489 - mean_absolute_error: 0.4984 Epoch 49/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2505 - mean_absolute_error: 0.4985 Epoch 50/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2527 - mean_absolute_error: 0.4983 Epoch 51/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2549 - mean_absolute_error: 0.4984 Epoch 52/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2549 - mean_absolute_error: 0.4986 Epoch 53/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2570 - mean_absolute_error: 0.4985 Epoch 54/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2579 - mean_absolute_error: 0.4987 Epoch 55/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2587 - mean_absolute_error: 0.4986 Epoch 56/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2602 - mean_absolute_error: 0.4986 Epoch 57/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2806 - mean_absolute_error: 0.4988 Epoch 58/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2712 - mean_absolute_error: 0.4988 Epoch 59/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2846 - mean_absolute_error: 0.4987 Epoch 60/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2883 - mean_absolute_error: 0.4989 Epoch 61/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2906 - mean_absolute_error: 0.4989 Epoch 62/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3006 - mean_absolute_error: 0.4988 Epoch 63/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3037 - mean_absolute_error: 0.4990 Epoch 64/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3025 - mean_absolute_error: 0.4988 Epoch 65/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3081 - mean_absolute_error: 0.4992 Epoch 66/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3059 - mean_absolute_error: 0.4989 Epoch 67/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2972 - mean_absolute_error: 0.4992 Epoch 68/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3078 - mean_absolute_error: 0.4990 Epoch 69/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3101 - mean_absolute_error: 0.4991 Epoch 70/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3110 - mean_absolute_error: 0.4992 Epoch 71/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3128 - mean_absolute_error: 0.4989 Epoch 72/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3129 - mean_absolute_error: 0.4991 Epoch 73/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3160 - mean_absolute_error: 0.4992 Epoch 74/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3149 - mean_absolute_error: 0.4987 Epoch 75/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3224 - mean_absolute_error: 0.4992 Epoch 76/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3152 - mean_absolute_error: 0.4993 Epoch 77/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3223 - mean_absolute_error: 0.4989 Epoch 78/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3266 - mean_absolute_error: 0.4992 Epoch 79/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3319 - mean_absolute_error: 0.4993 Epoch 80/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3290 - mean_absolute_error: 0.4992 Epoch 81/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3314 - mean_absolute_error: 0.4990 Epoch 82/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3329 - mean_absolute_error: 0.4992 Epoch 83/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3352 - mean_absolute_error: 0.4993 Epoch 84/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3362 - mean_absolute_error: 0.4991 Epoch 85/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3364 - mean_absolute_error: 0.4993 Epoch 86/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3368 - mean_absolute_error: 0.4993 Epoch 87/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3380 - mean_absolute_error: 0.4991 Epoch 88/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3399 - mean_absolute_error: 0.4993 Epoch 89/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3388 - mean_absolute_error: 0.4995 Epoch 90/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3425 - mean_absolute_error: 0.4993 Epoch 91/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3451 - mean_absolute_error: 0.4991 Epoch 92/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3455 - mean_absolute_error: 0.4993 Epoch 93/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3471 - mean_absolute_error: 0.4991 Epoch 94/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3477 - mean_absolute_error: 0.4992 Epoch 95/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3477 - mean_absolute_error: 0.4994 Epoch 96/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3481 - mean_absolute_error: 0.4992 Epoch 97/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3506 - mean_absolute_error: 0.4994 Epoch 98/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3514 - mean_absolute_error: 0.4993 Epoch 99/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3521 - mean_absolute_error: 0.4993 Epoch 100/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3540 - mean_absolute_error: 0.4993 Epoch 101/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3524 - mean_absolute_error: 0.4994 Epoch 102/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3523 - mean_absolute_error: 0.4992 Epoch 103/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3569 - mean_absolute_error: 0.4995 Epoch 104/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3582 - mean_absolute_error: 0.4994 Epoch 105/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3587 - mean_absolute_error: 0.4994 Epoch 106/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3600 - mean_absolute_error: 0.4994 Epoch 107/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3615 - mean_absolute_error: 0.4994 Epoch 108/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3608 - mean_absolute_error: 0.4995 Epoch 109/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3620 - mean_absolute_error: 0.4995 Epoch 110/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3643 - mean_absolute_error: 0.4994 Epoch 111/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3631 - mean_absolute_error: 0.4993 Epoch 112/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3680 - mean_absolute_error: 0.4994 Epoch 113/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3599 - mean_absolute_error: 0.4994 Epoch 114/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3682 - mean_absolute_error: 0.4994 1s - los Epoch 115/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3685 - mean_absolute_error: 0.4996 Epoch 116/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3700 - mean_absolute_error: 0.4996 Epoch 117/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3719 - mean_absolute_error: 0.4995 Epoch 118/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3716 - mean_absolute_error: 0.4996 Epoch 119/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3737 - mean_absolute_error: 0.4996 Epoch 120/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3742 - mean_absolute_error: 0.4996 Epoch 121/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3758 - mean_absolute_error: 0.4997 Epoch 122/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3763 - mean_absolute_error: 0.4997 Epoch 123/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3764 - mean_absolute_error: 0.4997 Epoch 124/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3780 - mean_absolute_error: 0.4998 Epoch 125/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3792 - mean_absolute_error: 0.4997 Epoch 126/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3778 - mean_absolute_error: 0.4996 Epoch 127/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3813 - mean_absolute_error: 0.4998 Epoch 128/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3791 - mean_absolute_error: 0.4998 Epoch 129/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3832 - mean_absolute_error: 0.4998 Epoch 130/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3835 - mean_absolute_error: 0.4999 Epoch 131/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3840 - mean_absolute_error: 0.4999 Epoch 132/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3855 - mean_absolute_error: 0.5000 Epoch 133/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3868 - mean_absolute_error: 0.5000 Epoch 134/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3864 - mean_absolute_error: 0.4999 Epoch 135/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3889 - mean_absolute_error: 0.5002 Epoch 136/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3885 - mean_absolute_error: 0.5000 Epoch 137/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3905 - mean_absolute_error: 0.5000 Epoch 138/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3908 - mean_absolute_error: 0.5000 Epoch 139/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3866 - mean_absolute_error: 0.5001 Epoch 140/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3951 - mean_absolute_error: 0.5000 Epoch 141/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3690 - mean_absolute_error: 0.5000 Epoch 142/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3653 - mean_absolute_error: 0.5002 Epoch 143/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3646 - mean_absolute_error: 0.5001 Epoch 144/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3674 - mean_absolute_error: 0.5003 Epoch 145/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3649 - mean_absolute_error: 0.5002 Epoch 146/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3642 - mean_absolute_error: 0.5005 Epoch 147/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3640 - mean_absolute_error: 0.5006 Epoch 148/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3648 - mean_absolute_error: 0.5005 Epoch 149/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3625 - mean_absolute_error: 0.5007 Epoch 150/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3630 - mean_absolute_error: 0.5008 Epoch 151/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3632 - mean_absolute_error: 0.5008 Epoch 152/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3619 - mean_absolute_error: 0.5008 Epoch 153/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3604 - mean_absolute_error: 0.5009 Epoch 154/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3617 - mean_absolute_error: 0.5009 Epoch 155/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3603 - mean_absolute_error: 0.5012 Epoch 156/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3597 - mean_absolute_error: 0.5011 Epoch 157/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3598 - mean_absolute_error: 0.5013 Epoch 158/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3594 - mean_absolute_error: 0.5012 Epoch 159/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3569 - mean_absolute_error: 0.5014 Epoch 160/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3594 - mean_absolute_error: 0.5013 Epoch 161/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3586 - mean_absolute_error: 0.5013 Epoch 162/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3578 - mean_absolute_error: 0.5015 Epoch 163/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3567 - mean_absolute_error: 0.5016 Epoch 164/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3563 - mean_absolute_error: 0.5015 Epoch 165/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3564 - mean_absolute_error: 0.5015 Epoch 166/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3561 - mean_absolute_error: 0.5016 Epoch 167/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3564 - mean_absolute_error: 0.5016 Epoch 168/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3561 - mean_absolute_error: 0.5016 Epoch 169/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3544 - mean_absolute_error: 0.5015 Epoch 170/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3560 - mean_absolute_error: 0.5015 Epoch 171/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3554 - mean_absolute_error: 0.5015 Epoch 172/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3525 - mean_absolute_error: 0.5016 Epoch 173/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3540 - mean_absolute_error: 0.5016 Epoch 174/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3533 - mean_absolute_error: 0.5017 Epoch 175/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3520 - mean_absolute_error: 0.5016 Epoch 176/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3496 - mean_absolute_error: 0.5015 1s - Epoch 177/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3478 - mean_absolute_error: 0.5017 Epoch 178/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3489 - mean_absolute_error: 0.5018 Epoch 179/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3470 - mean_absolute_error: 0.5017 Epoch 180/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3463 - mean_absolute_error: 0.5018 Epoch 181/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3473 - mean_absolute_error: 0.5016 Epoch 182/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3470 - mean_absolute_error: 0.5018 Epoch 183/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3467 - mean_absolute_error: 0.5020 Epoch 184/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3448 - mean_absolute_error: 0.5020 Epoch 185/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3441 - mean_absolute_error: 0.5017 Epoch 186/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3436 - mean_absolute_error: 0.5019 Epoch 187/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3452 - mean_absolute_error: 0.5019 Epoch 188/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3443 - mean_absolute_error: 0.5020 Epoch 189/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3447 - mean_absolute_error: 0.5019 Epoch 190/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3443 - mean_absolute_error: 0.5020 Epoch 191/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3476 - mean_absolute_error: 0.5018 Epoch 192/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3456 - mean_absolute_error: 0.5020 Epoch 193/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3442 - mean_absolute_error: 0.5017 Epoch 194/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3402 - mean_absolute_error: 0.5020 Epoch 195/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3406 - mean_absolute_error: 0.5021 Epoch 196/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3393 - mean_absolute_error: 0.5021 Epoch 197/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3383 - mean_absolute_error: 0.5020 Epoch 198/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3381 - mean_absolute_error: 0.5020 Epoch 199/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3390 - mean_absolute_error: 0.5022 Epoch 200/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3395 - mean_absolute_error: 0.5022 Epoch 201/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3438 - mean_absolute_error: 0.5022 Epoch 202/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3383 - mean_absolute_error: 0.5021 Epoch 203/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3378 - mean_absolute_error: 0.5021 Epoch 204/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3385 - mean_absolute_error: 0.5022 Epoch 205/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3372 - mean_absolute_error: 0.5023 Epoch 206/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3395 - mean_absolute_error: 0.5021 Epoch 207/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3352 - mean_absolute_error: 0.5022 Epoch 208/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3364 - mean_absolute_error: 0.5023 Epoch 209/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3390 - mean_absolute_error: 0.5022 Epoch 210/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3354 - mean_absolute_error: 0.5025 Epoch 211/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3392 - mean_absolute_error: 0.5021 Epoch 212/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3323 - mean_absolute_error: 0.5019 Epoch 213/250 64/64 [==============================] - 2s 36ms/step - loss: -2.3359 - mean_absolute_error: 0.5021 Epoch 214/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3325 - mean_absolute_error: 0.5021 Epoch 215/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3347 - mean_absolute_error: 0.5021 Epoch 216/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3329 - mean_absolute_error: 0.5020 Epoch 217/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3263 - mean_absolute_error: 0.5020 Epoch 218/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3304 - mean_absolute_error: 0.5021 Epoch 219/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3281 - mean_absolute_error: 0.5022 Epoch 220/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3319 - mean_absolute_error: 0.5020 Epoch 221/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3320 - mean_absolute_error: 0.5021 Epoch 222/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3288 - mean_absolute_error: 0.5021 Epoch 223/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3226 - mean_absolute_error: 0.5021 Epoch 224/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3252 - mean_absolute_error: 0.5023 Epoch 225/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3255 - mean_absolute_error: 0.5024 Epoch 226/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3268 - mean_absolute_error: 0.5025 Epoch 227/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3274 - mean_absolute_error: 0.5022 Epoch 228/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3169 - mean_absolute_error: 0.5019 Epoch 229/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3124 - mean_absolute_error: 0.5022 Epoch 230/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3198 - mean_absolute_error: 0.5024 Epoch 231/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3167 - mean_absolute_error: 0.5023 Epoch 232/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3097 - mean_absolute_error: 0.5021 Epoch 233/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3167 - mean_absolute_error: 0.5023 Epoch 234/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3117 - mean_absolute_error: 0.5022 Epoch 235/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3089 - mean_absolute_error: 0.5020 Epoch 236/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3111 - mean_absolute_error: 0.5024 Epoch 237/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3063 - mean_absolute_error: 0.5019 Epoch 238/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3000 - mean_absolute_error: 0.5022 Epoch 239/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3077 - mean_absolute_error: 0.5021 Epoch 240/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2911 - mean_absolute_error: 0.5020 Epoch 241/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2986 - mean_absolute_error: 0.5023 Epoch 242/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2954 - mean_absolute_error: 0.5023 Epoch 243/250 64/64 [==============================] - 2s 32ms/step - loss: -2.2865 - mean_absolute_error: 0.5021 Epoch 244/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2913 - mean_absolute_error: 0.5023 Epoch 245/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2985 - mean_absolute_error: 0.5022 Epoch 246/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2856 - mean_absolute_error: 0.5019 Epoch 247/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2855 - mean_absolute_error: 0.5023 Epoch 248/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3016 - mean_absolute_error: 0.5025 Epoch 249/250 64/64 [==============================] - 2s 32ms/step - loss: -2.2883 - mean_absolute_error: 0.5023 Epoch 250/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2831 - mean_absolute_error: 0.5018 Training model 3 of 5. Epoch 1/250 64/64 [==============================] - 6s 35ms/step - loss: 6.3971 - mean_absolute_error: 0.5198 Epoch 2/250 64/64 [==============================] - 3s 49ms/step - loss: -1.2546 - mean_absolute_error: 0.4627 Epoch 3/250 64/64 [==============================] - 3s 42ms/step - loss: -1.5099 - mean_absolute_error: 0.4689 Epoch 4/250 64/64 [==============================] - 3s 40ms/step - loss: -1.5950 - mean_absolute_error: 0.4702 Epoch 5/250 64/64 [==============================] - 2s 34ms/step - loss: -1.6479 - mean_absolute_error: 0.4732 Epoch 6/250 64/64 [==============================] - 2s 34ms/step - loss: -1.6875 - mean_absolute_error: 0.4741 Epoch 7/250 64/64 [==============================] - 2s 37ms/step - loss: -1.7199 - mean_absolute_error: 0.4761 Epoch 8/250 64/64 [==============================] - 2s 36ms/step - loss: -1.7513 - mean_absolute_error: 0.4773 Epoch 9/250 64/64 [==============================] - 2s 35ms/step - loss: -1.7726 - mean_absolute_error: 0.4790 Epoch 10/250 64/64 [==============================] - 2s 33ms/step - loss: -1.7932 - mean_absolute_error: 0.4800 Epoch 11/250 64/64 [==============================] - 2s 33ms/step - loss: -1.8125 - mean_absolute_error: 0.4807 Epoch 12/250 64/64 [==============================] - 2s 34ms/step - loss: -1.8290 - mean_absolute_error: 0.4815 Epoch 13/250 64/64 [==============================] - 2s 35ms/step - loss: -1.8444 - mean_absolute_error: 0.4820 Epoch 14/250 64/64 [==============================] - 2s 34ms/step - loss: -1.8589 - mean_absolute_error: 0.4827 Epoch 15/250 64/64 [==============================] - 2s 35ms/step - loss: -1.8715 - mean_absolute_error: 0.4831 Epoch 16/250 64/64 [==============================] - 2s 35ms/step - loss: -1.8841 - mean_absolute_error: 0.4840 Epoch 17/250 64/64 [==============================] - 2s 34ms/step - loss: -1.9007 - mean_absolute_error: 0.4843 Epoch 18/250 64/64 [==============================] - 2s 34ms/step - loss: -1.9151 - mean_absolute_error: 0.4853 Epoch 19/250 64/64 [==============================] - 2s 34ms/step - loss: -1.9251 - mean_absolute_error: 0.4858 Epoch 20/250 64/64 [==============================] - 2s 37ms/step - loss: -1.9371 - mean_absolute_error: 0.4860 Epoch 21/250 64/64 [==============================] - 2s 36ms/step - loss: -1.9483 - mean_absolute_error: 0.4865 Epoch 22/250 64/64 [==============================] - 2s 35ms/step - loss: -1.9603 - mean_absolute_error: 0.4873 Epoch 23/250 64/64 [==============================] - 2s 34ms/step - loss: -1.9675 - mean_absolute_error: 0.4873 Epoch 24/250 64/64 [==============================] - 2s 34ms/step - loss: -1.9798 - mean_absolute_error: 0.4881 Epoch 25/250 64/64 [==============================] - 2s 34ms/step - loss: -1.9887 - mean_absolute_error: 0.4884 Epoch 26/250 64/64 [==============================] - 2s 34ms/step - loss: -1.9953 - mean_absolute_error: 0.4886 Epoch 27/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0012 - mean_absolute_error: 0.4889 Epoch 28/250 64/64 [==============================] - 2s 33ms/step - loss: -2.0105 - mean_absolute_error: 0.4892 Epoch 29/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0172 - mean_absolute_error: 0.4892 Epoch 30/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0224 - mean_absolute_error: 0.4895 Epoch 31/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0294 - mean_absolute_error: 0.4898 Epoch 32/250 64/64 [==============================] - 2s 35ms/step - loss: -2.0352 - mean_absolute_error: 0.4899 Epoch 33/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0429 - mean_absolute_error: 0.4900 Epoch 34/250 64/64 [==============================] - 2s 36ms/step - loss: -2.0465 - mean_absolute_error: 0.4902 Epoch 35/250 64/64 [==============================] - 2s 37ms/step - loss: -2.0514 - mean_absolute_error: 0.4903 Epoch 36/250 64/64 [==============================] - 2s 37ms/step - loss: -2.0592 - mean_absolute_error: 0.4902 Epoch 37/250 64/64 [==============================] - 2s 35ms/step - loss: -2.0637 - mean_absolute_error: 0.4906 Epoch 38/250 64/64 [==============================] - 2s 36ms/step - loss: -2.0676 - mean_absolute_error: 0.4902 Epoch 39/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0724 - mean_absolute_error: 0.4904 Epoch 40/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0739 - mean_absolute_error: 0.4904 Epoch 41/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0776 - mean_absolute_error: 0.4905 Epoch 42/250 64/64 [==============================] - 2s 35ms/step - loss: -2.0817 - mean_absolute_error: 0.4906 Epoch 43/250 64/64 [==============================] - 2s 33ms/step - loss: -2.0854 - mean_absolute_error: 0.4905 Epoch 44/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0875 - mean_absolute_error: 0.4907 Epoch 45/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0914 - mean_absolute_error: 0.4906 Epoch 46/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0949 - mean_absolute_error: 0.4909 Epoch 47/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0956 - mean_absolute_error: 0.4903 Epoch 48/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0975 - mean_absolute_error: 0.4909 Epoch 49/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1008 - mean_absolute_error: 0.4907 Epoch 50/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1020 - mean_absolute_error: 0.4908 Epoch 51/250 64/64 [==============================] - 2s 36ms/step - loss: -2.1073 - mean_absolute_error: 0.4909 Epoch 52/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1078 - mean_absolute_error: 0.4908 Epoch 53/250 64/64 [==============================] - 2s 37ms/step - loss: -2.1132 - mean_absolute_error: 0.4909 Epoch 54/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1132 - mean_absolute_error: 0.4911 Epoch 55/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1176 - mean_absolute_error: 0.4910 Epoch 56/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1181 - mean_absolute_error: 0.4910 Epoch 57/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1165 - mean_absolute_error: 0.4909 Epoch 58/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1191 - mean_absolute_error: 0.4909 Epoch 59/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1218 - mean_absolute_error: 0.4911 Epoch 60/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1235 - mean_absolute_error: 0.4912 Epoch 61/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1248 - mean_absolute_error: 0.4912 Epoch 62/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1256 - mean_absolute_error: 0.4910 Epoch 63/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1290 - mean_absolute_error: 0.4913 Epoch 64/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1331 - mean_absolute_error: 0.4912 Epoch 65/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1348 - mean_absolute_error: 0.4913 Epoch 66/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1351 - mean_absolute_error: 0.4911 Epoch 67/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1347 - mean_absolute_error: 0.4912 Epoch 68/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1349 - mean_absolute_error: 0.4914 Epoch 69/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1384 - mean_absolute_error: 0.4914 Epoch 70/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1392 - mean_absolute_error: 0.4914 Epoch 71/250 64/64 [==============================] - 2s 36ms/step - loss: -2.1413 - mean_absolute_error: 0.4913 Epoch 72/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1423 - mean_absolute_error: 0.4913 Epoch 73/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1458 - mean_absolute_error: 0.4913 Epoch 74/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1445 - mean_absolute_error: 0.4912 Epoch 75/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1462 - mean_absolute_error: 0.4913 Epoch 76/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1464 - mean_absolute_error: 0.4914 Epoch 77/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1472 - mean_absolute_error: 0.4912 Epoch 78/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1501 - mean_absolute_error: 0.4914 Epoch 79/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1513 - mean_absolute_error: 0.4914 Epoch 80/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1527 - mean_absolute_error: 0.4913 Epoch 81/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1561 - mean_absolute_error: 0.4915 Epoch 82/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1548 - mean_absolute_error: 0.4914 Epoch 83/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1525 - mean_absolute_error: 0.4911 Epoch 84/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1544 - mean_absolute_error: 0.4912 Epoch 85/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1578 - mean_absolute_error: 0.4913 Epoch 86/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1596 - mean_absolute_error: 0.4914 Epoch 87/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1613 - mean_absolute_error: 0.4916 Epoch 88/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1596 - mean_absolute_error: 0.4916 Epoch 89/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1599 - mean_absolute_error: 0.4914 Epoch 90/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1610 - mean_absolute_error: 0.4916 Epoch 91/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1628 - mean_absolute_error: 0.4915 Epoch 92/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1647 - mean_absolute_error: 0.4915 Epoch 93/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1670 - mean_absolute_error: 0.4916 Epoch 94/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1670 - mean_absolute_error: 0.4916 Epoch 95/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1668 - mean_absolute_error: 0.4916 Epoch 96/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1653 - mean_absolute_error: 0.4916 Epoch 97/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1681 - mean_absolute_error: 0.4917 Epoch 98/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1692 - mean_absolute_error: 0.4917 Epoch 99/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1703 - mean_absolute_error: 0.4918 Epoch 100/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1724 - mean_absolute_error: 0.4917 Epoch 101/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1749 - mean_absolute_error: 0.4918 Epoch 102/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1751 - mean_absolute_error: 0.4919 Epoch 103/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1766 - mean_absolute_error: 0.4917 Epoch 104/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1753 - mean_absolute_error: 0.4917 Epoch 105/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1735 - mean_absolute_error: 0.4917 Epoch 106/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1750 - mean_absolute_error: 0.4918 Epoch 107/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1754 - mean_absolute_error: 0.4918 Epoch 108/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1805 - mean_absolute_error: 0.4918 Epoch 109/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1803 - mean_absolute_error: 0.4919 Epoch 110/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1814 - mean_absolute_error: 0.4917 Epoch 111/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1782 - mean_absolute_error: 0.4919 Epoch 112/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1797 - mean_absolute_error: 0.4917 Epoch 113/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1820 - mean_absolute_error: 0.4918 1s - loss: -2.18 Epoch 114/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1842 - mean_absolute_error: 0.4919 Epoch 115/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1851 - mean_absolute_error: 0.4921 Epoch 116/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1860 - mean_absolute_error: 0.4918 Epoch 117/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1878 - mean_absolute_error: 0.4918 Epoch 118/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1856 - mean_absolute_error: 0.4919 Epoch 119/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1858 - mean_absolute_error: 0.4919 Epoch 120/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1862 - mean_absolute_error: 0.4921 Epoch 121/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1869 - mean_absolute_error: 0.4920 Epoch 122/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1908 - mean_absolute_error: 0.4920 Epoch 123/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1907 - mean_absolute_error: 0.4919 Epoch 124/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1902 - mean_absolute_error: 0.4923 Epoch 125/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1911 - mean_absolute_error: 0.4921 Epoch 126/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1896 - mean_absolute_error: 0.4919 Epoch 127/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1924 - mean_absolute_error: 0.4919 Epoch 128/250 64/64 [==============================] - 2s 36ms/step - loss: -2.1928 - mean_absolute_error: 0.4920 Epoch 129/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1955 - mean_absolute_error: 0.4922 Epoch 130/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1928 - mean_absolute_error: 0.4920 Epoch 131/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1919 - mean_absolute_error: 0.4920 Epoch 132/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1955 - mean_absolute_error: 0.4919 Epoch 133/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1971 - mean_absolute_error: 0.4921 Epoch 134/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1968 - mean_absolute_error: 0.4921 Epoch 135/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1990 - mean_absolute_error: 0.4922 Epoch 136/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1986 - mean_absolute_error: 0.4922 Epoch 137/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1963 - mean_absolute_error: 0.4922 Epoch 138/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1949 - mean_absolute_error: 0.4919 Epoch 139/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1997 - mean_absolute_error: 0.4919 Epoch 140/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2006 - mean_absolute_error: 0.4922 Epoch 141/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2018 - mean_absolute_error: 0.4922 Epoch 142/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2007 - mean_absolute_error: 0.4922 Epoch 143/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2004 - mean_absolute_error: 0.4921 Epoch 144/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1996 - mean_absolute_error: 0.4921 Epoch 145/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2034 - mean_absolute_error: 0.4922 Epoch 146/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2041 - mean_absolute_error: 0.4920 Epoch 147/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2042 - mean_absolute_error: 0.4922 Epoch 148/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2012 - mean_absolute_error: 0.4919 Epoch 149/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2030 - mean_absolute_error: 0.4922 Epoch 150/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2022 - mean_absolute_error: 0.4922 Epoch 151/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2033 - mean_absolute_error: 0.4921 Epoch 152/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2038 - mean_absolute_error: 0.4922 Epoch 153/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2070 - mean_absolute_error: 0.4922 Epoch 154/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2075 - mean_absolute_error: 0.4923 Epoch 155/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2081 - mean_absolute_error: 0.4921 Epoch 156/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2097 - mean_absolute_error: 0.4922 Epoch 157/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2091 - mean_absolute_error: 0.4923 Epoch 158/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2089 - mean_absolute_error: 0.4923 Epoch 159/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2071 - mean_absolute_error: 0.4923 Epoch 160/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2079 - mean_absolute_error: 0.4924 Epoch 161/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2063 - mean_absolute_error: 0.4922 Epoch 162/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2114 - mean_absolute_error: 0.4923 Epoch 163/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2091 - mean_absolute_error: 0.4924 Epoch 164/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2095 - mean_absolute_error: 0.4922 Epoch 165/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2107 - mean_absolute_error: 0.4923 Epoch 166/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2104 - mean_absolute_error: 0.4920 Epoch 167/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2101 - mean_absolute_error: 0.4922 Epoch 168/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2113 - mean_absolute_error: 0.4920 Epoch 169/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2129 - mean_absolute_error: 0.4923 Epoch 170/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2155 - mean_absolute_error: 0.4923 Epoch 171/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2120 - mean_absolute_error: 0.4923 Epoch 172/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2117 - mean_absolute_error: 0.4922 Epoch 173/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2119 - mean_absolute_error: 0.4924 Epoch 174/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2131 - mean_absolute_error: 0.4924 Epoch 175/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2149 - mean_absolute_error: 0.4926 Epoch 176/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2164 - mean_absolute_error: 0.4925 Epoch 177/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2169 - mean_absolute_error: 0.4924 Epoch 178/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2167 - mean_absolute_error: 0.4923 Epoch 179/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2154 - mean_absolute_error: 0.4924 Epoch 180/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2156 - mean_absolute_error: 0.4926 Epoch 181/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2154 - mean_absolute_error: 0.4925 Epoch 182/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2161 - mean_absolute_error: 0.4925 Epoch 183/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2182 - mean_absolute_error: 0.4925 Epoch 184/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2202 - mean_absolute_error: 0.4925 Epoch 185/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2219 - mean_absolute_error: 0.4924 Epoch 186/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2208 - mean_absolute_error: 0.4926 Epoch 187/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2216 - mean_absolute_error: 0.4926 Epoch 188/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2198 - mean_absolute_error: 0.4925 Epoch 189/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2188 - mean_absolute_error: 0.4925 Epoch 190/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2215 - mean_absolute_error: 0.4925 Epoch 191/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2228 - mean_absolute_error: 0.4923 Epoch 192/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2237 - mean_absolute_error: 0.4927 Epoch 193/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2227 - mean_absolute_error: 0.4924 Epoch 194/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2217 - mean_absolute_error: 0.4924 Epoch 195/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2237 - mean_absolute_error: 0.4924 Epoch 196/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2258 - mean_absolute_error: 0.4923 Epoch 197/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2256 - mean_absolute_error: 0.4926 Epoch 198/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2225 - mean_absolute_error: 0.4924 Epoch 199/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2205 - mean_absolute_error: 0.4925 Epoch 200/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2229 - mean_absolute_error: 0.4926 Epoch 201/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2231 - mean_absolute_error: 0.4926 Epoch 202/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2249 - mean_absolute_error: 0.4926 Epoch 203/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2279 - mean_absolute_error: 0.4926 Epoch 204/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2283 - mean_absolute_error: 0.4927 Epoch 205/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2257 - mean_absolute_error: 0.4926 Epoch 206/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2279 - mean_absolute_error: 0.4927 Epoch 207/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2260 - mean_absolute_error: 0.4927 Epoch 208/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2250 - mean_absolute_error: 0.4926 Epoch 209/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2273 - mean_absolute_error: 0.4925 Epoch 210/250 64/64 [==============================] - 2s 36ms/step - loss: -2.2309 - mean_absolute_error: 0.4927 Epoch 211/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2288 - mean_absolute_error: 0.4926 Epoch 212/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2264 - mean_absolute_error: 0.4927 Epoch 213/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2270 - mean_absolute_error: 0.4927 Epoch 214/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2287 - mean_absolute_error: 0.4926 Epoch 215/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2315 - mean_absolute_error: 0.4928 Epoch 216/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2310 - mean_absolute_error: 0.4928 Epoch 217/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2305 - mean_absolute_error: 0.4929 Epoch 218/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2308 - mean_absolute_error: 0.4928 Epoch 219/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2306 - mean_absolute_error: 0.4928 Epoch 220/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2316 - mean_absolute_error: 0.4930 Epoch 221/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2323 - mean_absolute_error: 0.4930 Epoch 222/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2310 - mean_absolute_error: 0.4929 Epoch 223/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2313 - mean_absolute_error: 0.4931 Epoch 224/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2323 - mean_absolute_error: 0.4931 Epoch 225/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2326 - mean_absolute_error: 0.4932 Epoch 226/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2362 - mean_absolute_error: 0.4932 Epoch 227/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2360 - mean_absolute_error: 0.4930 Epoch 228/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2336 - mean_absolute_error: 0.4931 Epoch 229/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2310 - mean_absolute_error: 0.4930 Epoch 230/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2358 - mean_absolute_error: 0.4933 Epoch 231/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2372 - mean_absolute_error: 0.4932 Epoch 232/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2363 - mean_absolute_error: 0.4932 Epoch 233/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2329 - mean_absolute_error: 0.4933 Epoch 234/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2321 - mean_absolute_error: 0.4932 Epoch 235/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2371 - mean_absolute_error: 0.4933 Epoch 236/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2363 - mean_absolute_error: 0.4933 Epoch 237/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2381 - mean_absolute_error: 0.4935 Epoch 238/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2386 - mean_absolute_error: 0.4935 Epoch 239/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2370 - mean_absolute_error: 0.4934 Epoch 240/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2352 - mean_absolute_error: 0.4935 Epoch 241/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2347 - mean_absolute_error: 0.4936 Epoch 242/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2389 - mean_absolute_error: 0.4935 Epoch 243/250 64/64 [==============================] - 2s 36ms/step - loss: -2.2401 - mean_absolute_error: 0.4937 Epoch 244/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2404 - mean_absolute_error: 0.4937 Epoch 245/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2383 - mean_absolute_error: 0.4936 Epoch 246/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2361 - mean_absolute_error: 0.4937 Epoch 247/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2377 - mean_absolute_error: 0.4937 Epoch 248/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2399 - mean_absolute_error: 0.4938 Epoch 249/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2421 - mean_absolute_error: 0.4939 Epoch 250/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2412 - mean_absolute_error: 0.4937 Training model 4 of 5. Epoch 1/250 64/64 [==============================] - 19s 104ms/step - loss: -0.3403 - mean_absolute_error: 0.5312 Epoch 2/250 64/64 [==============================] - 2s 33ms/step - loss: -1.3326 - mean_absolute_error: 0.4744 Epoch 3/250 64/64 [==============================] - 2s 34ms/step - loss: -1.6183 - mean_absolute_error: 0.4806 Epoch 4/250 64/64 [==============================] - 2s 34ms/step - loss: -1.7162 - mean_absolute_error: 0.4833 Epoch 5/250 64/64 [==============================] - 2s 33ms/step - loss: -1.7854 - mean_absolute_error: 0.4858 Epoch 6/250 64/64 [==============================] - 2s 34ms/step - loss: -1.8408 - mean_absolute_error: 0.4888 Epoch 7/250 64/64 [==============================] - 2s 34ms/step - loss: -1.8918 - mean_absolute_error: 0.4910 Epoch 8/250 64/64 [==============================] - 2s 34ms/step - loss: -1.9259 - mean_absolute_error: 0.4924 Epoch 9/250 64/64 [==============================] - 2s 35ms/step - loss: -1.9533 - mean_absolute_error: 0.4930 Epoch 10/250 64/64 [==============================] - 2s 33ms/step - loss: -1.9753 - mean_absolute_error: 0.4937 Epoch 11/250 64/64 [==============================] - 3s 40ms/step - loss: -1.9980 - mean_absolute_error: 0.4943 Epoch 12/250 64/64 [==============================] - 2s 35ms/step - loss: -2.0140 - mean_absolute_error: 0.4948 Epoch 13/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0304 - mean_absolute_error: 0.4955 Epoch 14/250 64/64 [==============================] - 2s 35ms/step - loss: -2.0426 - mean_absolute_error: 0.4957 Epoch 15/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0561 - mean_absolute_error: 0.4962 Epoch 16/250 64/64 [==============================] - 2s 33ms/step - loss: -2.0663 - mean_absolute_error: 0.4966 Epoch 17/250 64/64 [==============================] - 2s 33ms/step - loss: -2.0751 - mean_absolute_error: 0.4967 Epoch 18/250 64/64 [==============================] - 2s 33ms/step - loss: -2.0840 - mean_absolute_error: 0.4970 Epoch 19/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0928 - mean_absolute_error: 0.4966 Epoch 20/250 64/64 [==============================] - 2s 35ms/step - loss: -2.0996 - mean_absolute_error: 0.4973 Epoch 21/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1064 - mean_absolute_error: 0.4974 Epoch 22/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1130 - mean_absolute_error: 0.4976 Epoch 23/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1181 - mean_absolute_error: 0.4980 Epoch 24/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1242 - mean_absolute_error: 0.4979 Epoch 25/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1298 - mean_absolute_error: 0.4984 Epoch 26/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1355 - mean_absolute_error: 0.4984 Epoch 27/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1397 - mean_absolute_error: 0.4986 Epoch 28/250 64/64 [==============================] - 2s 38ms/step - loss: -2.1440 - mean_absolute_error: 0.4987 Epoch 29/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1490 - mean_absolute_error: 0.4986 Epoch 30/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1530 - mean_absolute_error: 0.4991 Epoch 31/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1560 - mean_absolute_error: 0.4990 Epoch 32/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1612 - mean_absolute_error: 0.4992 Epoch 33/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1658 - mean_absolute_error: 0.4994 Epoch 34/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1694 - mean_absolute_error: 0.4994 Epoch 35/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1723 - mean_absolute_error: 0.4996 Epoch 36/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1742 - mean_absolute_error: 0.4998 Epoch 37/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1779 - mean_absolute_error: 0.5005 Epoch 38/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1864 - mean_absolute_error: 0.5002 Epoch 39/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1901 - mean_absolute_error: 0.5003 Epoch 40/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1941 - mean_absolute_error: 0.5005 Epoch 41/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1979 - mean_absolute_error: 0.5003 Epoch 42/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2002 - mean_absolute_error: 0.5004 Epoch 43/250 64/64 [==============================] - 2s 36ms/step - loss: -2.2040 - mean_absolute_error: 0.5005 Epoch 44/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2062 - mean_absolute_error: 0.5007 Epoch 45/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2104 - mean_absolute_error: 0.5006 Epoch 46/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2133 - mean_absolute_error: 0.5006 Epoch 47/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2154 - mean_absolute_error: 0.5008 Epoch 48/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2167 - mean_absolute_error: 0.5009 Epoch 49/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2211 - mean_absolute_error: 0.5011 Epoch 50/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2227 - mean_absolute_error: 0.5010 Epoch 51/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2255 - mean_absolute_error: 0.5011 Epoch 52/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2281 - mean_absolute_error: 0.5012 Epoch 53/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2289 - mean_absolute_error: 0.5013 Epoch 54/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2331 - mean_absolute_error: 0.5014 Epoch 55/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2351 - mean_absolute_error: 0.5015 Epoch 56/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2374 - mean_absolute_error: 0.5014 Epoch 57/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2393 - mean_absolute_error: 0.5017 Epoch 58/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2415 - mean_absolute_error: 0.5017 Epoch 59/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2431 - mean_absolute_error: 0.5015 Epoch 60/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2458 - mean_absolute_error: 0.5019 Epoch 61/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2485 - mean_absolute_error: 0.5019 Epoch 62/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2495 - mean_absolute_error: 0.5019 Epoch 63/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2527 - mean_absolute_error: 0.5021 Epoch 64/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2527 - mean_absolute_error: 0.5022 Epoch 65/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2539 - mean_absolute_error: 0.5020 Epoch 66/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2587 - mean_absolute_error: 0.5022 Epoch 67/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2592 - mean_absolute_error: 0.5023 Epoch 68/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2591 - mean_absolute_error: 0.5024 Epoch 69/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2607 - mean_absolute_error: 0.5023 Epoch 70/250 64/64 [==============================] - 2s 36ms/step - loss: -2.2636 - mean_absolute_error: 0.5024 Epoch 71/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2661 - mean_absolute_error: 0.5025 Epoch 72/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2655 - mean_absolute_error: 0.5024 Epoch 73/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2675 - mean_absolute_error: 0.5025 Epoch 74/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2694 - mean_absolute_error: 0.5026 Epoch 75/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2714 - mean_absolute_error: 0.5027 Epoch 76/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2720 - mean_absolute_error: 0.5027 Epoch 77/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2736 - mean_absolute_error: 0.5027 Epoch 78/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2761 - mean_absolute_error: 0.5027 Epoch 79/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2793 - mean_absolute_error: 0.5028 Epoch 80/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2794 - mean_absolute_error: 0.5029 Epoch 81/250 64/64 [==============================] - 2s 36ms/step - loss: -2.2821 - mean_absolute_error: 0.5029 Epoch 82/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2833 - mean_absolute_error: 0.5028 Epoch 83/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2835 - mean_absolute_error: 0.5030 Epoch 84/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2850 - mean_absolute_error: 0.5029 Epoch 85/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2870 - mean_absolute_error: 0.5029 Epoch 86/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2881 - mean_absolute_error: 0.5029 Epoch 87/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2880 - mean_absolute_error: 0.5030 Epoch 88/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2888 - mean_absolute_error: 0.5029 Epoch 89/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2903 - mean_absolute_error: 0.5030 Epoch 90/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2915 - mean_absolute_error: 0.5032 Epoch 91/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2899 - mean_absolute_error: 0.5029 Epoch 92/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2931 - mean_absolute_error: 0.5028 Epoch 93/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2937 - mean_absolute_error: 0.5029 Epoch 94/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2958 - mean_absolute_error: 0.5031 Epoch 95/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2973 - mean_absolute_error: 0.5030 Epoch 96/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2972 - mean_absolute_error: 0.5030 Epoch 97/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2971 - mean_absolute_error: 0.5032 Epoch 98/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2913 - mean_absolute_error: 0.5030 Epoch 99/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2829 - mean_absolute_error: 0.5028 Epoch 100/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2985 - mean_absolute_error: 0.5033 Epoch 101/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3036 - mean_absolute_error: 0.5031 Epoch 102/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3066 - mean_absolute_error: 0.5031 Epoch 103/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3052 - mean_absolute_error: 0.5031 Epoch 104/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3048 - mean_absolute_error: 0.5030 Epoch 105/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3074 - mean_absolute_error: 0.5032 Epoch 106/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3099 - mean_absolute_error: 0.5031 Epoch 107/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3092 - mean_absolute_error: 0.5031 Epoch 108/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3105 - mean_absolute_error: 0.5032 Epoch 109/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3120 - mean_absolute_error: 0.5033 Epoch 110/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3136 - mean_absolute_error: 0.5032 Epoch 111/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3125 - mean_absolute_error: 0.5032 Epoch 112/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3138 - mean_absolute_error: 0.5033 Epoch 113/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3156 - mean_absolute_error: 0.5032 Epoch 114/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3161 - mean_absolute_error: 0.5031 Epoch 115/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3162 - mean_absolute_error: 0.5030 Epoch 116/250 64/64 [==============================] - 2s 36ms/step - loss: -2.3174 - mean_absolute_error: 0.5030 Epoch 117/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3182 - mean_absolute_error: 0.5031 Epoch 118/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3203 - mean_absolute_error: 0.5031 Epoch 119/250 64/64 [==============================] - 2s 36ms/step - loss: -2.3210 - mean_absolute_error: 0.5032 Epoch 120/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3201 - mean_absolute_error: 0.5029 Epoch 121/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3215 - mean_absolute_error: 0.5031 Epoch 122/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3243 - mean_absolute_error: 0.5031 Epoch 123/250 64/64 [==============================] - 2s 36ms/step - loss: -2.3252 - mean_absolute_error: 0.5030 Epoch 124/250 64/64 [==============================] - 2s 36ms/step - loss: -2.3231 - mean_absolute_error: 0.5030 Epoch 125/250 64/64 [==============================] - 2s 36ms/step - loss: -2.3238 - mean_absolute_error: 0.5031 Epoch 126/250 64/64 [==============================] - 2s 36ms/step - loss: -2.3268 - mean_absolute_error: 0.5030 Epoch 127/250 64/64 [==============================] - 2s 36ms/step - loss: -2.3268 - mean_absolute_error: 0.5030 Epoch 128/250 64/64 [==============================] - 2s 36ms/step - loss: -2.3278 - mean_absolute_error: 0.5030 Epoch 129/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3260 - mean_absolute_error: 0.5032 Epoch 130/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3276 - mean_absolute_error: 0.5032 Epoch 131/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3291 - mean_absolute_error: 0.5029 Epoch 132/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3309 - mean_absolute_error: 0.5030 Epoch 133/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3309 - mean_absolute_error: 0.5031 Epoch 134/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3319 - mean_absolute_error: 0.5031 Epoch 135/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3302 - mean_absolute_error: 0.5032 Epoch 136/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3339 - mean_absolute_error: 0.5034 Epoch 137/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3347 - mean_absolute_error: 0.5032 Epoch 138/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3351 - mean_absolute_error: 0.5032 Epoch 139/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3337 - mean_absolute_error: 0.5030 Epoch 140/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3352 - mean_absolute_error: 0.5032 Epoch 141/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3372 - mean_absolute_error: 0.5033 Epoch 142/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3377 - mean_absolute_error: 0.5031 Epoch 143/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3371 - mean_absolute_error: 0.5032 Epoch 144/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3373 - mean_absolute_error: 0.5034 Epoch 145/250 64/64 [==============================] - 2s 36ms/step - loss: -2.3413 - mean_absolute_error: 0.5033 Epoch 146/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3408 - mean_absolute_error: 0.5033 Epoch 147/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3415 - mean_absolute_error: 0.5033 Epoch 148/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3412 - mean_absolute_error: 0.5033 Epoch 149/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3431 - mean_absolute_error: 0.5033 Epoch 150/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3438 - mean_absolute_error: 0.5032 Epoch 151/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3443 - mean_absolute_error: 0.5035 Epoch 152/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3442 - mean_absolute_error: 0.5034 Epoch 153/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3452 - mean_absolute_error: 0.5032 Epoch 154/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3439 - mean_absolute_error: 0.5034 Epoch 155/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3469 - mean_absolute_error: 0.5033 Epoch 156/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3482 - mean_absolute_error: 0.5034 Epoch 157/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3480 - mean_absolute_error: 0.5033 Epoch 158/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3472 - mean_absolute_error: 0.5033 Epoch 159/250 64/64 [==============================] - 2s 36ms/step - loss: -2.3496 - mean_absolute_error: 0.5034 Epoch 160/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3491 - mean_absolute_error: 0.5035 Epoch 161/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3505 - mean_absolute_error: 0.5035 Epoch 162/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3498 - mean_absolute_error: 0.5035 Epoch 163/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3495 - mean_absolute_error: 0.5035 Epoch 164/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3513 - mean_absolute_error: 0.5033 Epoch 165/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3518 - mean_absolute_error: 0.5035 Epoch 166/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3536 - mean_absolute_error: 0.5034 Epoch 167/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3531 - mean_absolute_error: 0.5036 Epoch 168/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3529 - mean_absolute_error: 0.5034 Epoch 169/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3549 - mean_absolute_error: 0.5035 Epoch 170/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3553 - mean_absolute_error: 0.5034 Epoch 171/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3546 - mean_absolute_error: 0.5036 Epoch 172/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3545 - mean_absolute_error: 0.5035 Epoch 173/250 64/64 [==============================] - 2s 36ms/step - loss: -2.3580 - mean_absolute_error: 0.5036 Epoch 174/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3586 - mean_absolute_error: 0.5036 Epoch 175/250 64/64 [==============================] - 2s 36ms/step - loss: -2.3566 - mean_absolute_error: 0.5037 Epoch 176/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3571 - mean_absolute_error: 0.5036 Epoch 177/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3578 - mean_absolute_error: 0.5038 Epoch 178/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3594 - mean_absolute_error: 0.5035 Epoch 179/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3587 - mean_absolute_error: 0.5036 Epoch 180/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3592 - mean_absolute_error: 0.5036 Epoch 181/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3603 - mean_absolute_error: 0.5037 Epoch 182/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3624 - mean_absolute_error: 0.5037 Epoch 183/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3626 - mean_absolute_error: 0.5036 Epoch 184/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3597 - mean_absolute_error: 0.5037 Epoch 185/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3612 - mean_absolute_error: 0.5036 Epoch 186/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3637 - mean_absolute_error: 0.5037 Epoch 187/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3652 - mean_absolute_error: 0.5037 Epoch 188/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3613 - mean_absolute_error: 0.5035 Epoch 189/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3611 - mean_absolute_error: 0.5038 Epoch 190/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3625 - mean_absolute_error: 0.5037 Epoch 191/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3657 - mean_absolute_error: 0.5037 Epoch 192/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3639 - mean_absolute_error: 0.5036 Epoch 193/250 64/64 [==============================] - 2s 36ms/step - loss: -2.3626 - mean_absolute_error: 0.5043 Epoch 194/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3642 - mean_absolute_error: 0.5043 Epoch 195/250 64/64 [==============================] - 2s 36ms/step - loss: -2.3659 - mean_absolute_error: 0.5043 Epoch 196/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3656 - mean_absolute_error: 0.5041 Epoch 197/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3640 - mean_absolute_error: 0.5040 Epoch 198/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3649 - mean_absolute_error: 0.5038 Epoch 199/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3667 - mean_absolute_error: 0.5038 Epoch 200/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3673 - mean_absolute_error: 0.5038 Epoch 201/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3662 - mean_absolute_error: 0.5037 Epoch 202/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3668 - mean_absolute_error: 0.5037 Epoch 203/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3709 - mean_absolute_error: 0.5038 Epoch 204/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3691 - mean_absolute_error: 0.5041 Epoch 205/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3680 - mean_absolute_error: 0.5039 Epoch 206/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3670 - mean_absolute_error: 0.5038 Epoch 207/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3669 - mean_absolute_error: 0.5038 Epoch 208/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3696 - mean_absolute_error: 0.5037 Epoch 209/250 64/64 [==============================] - 2s 36ms/step - loss: -2.3701 - mean_absolute_error: 0.5038 Epoch 210/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3689 - mean_absolute_error: 0.5042 Epoch 211/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3678 - mean_absolute_error: 0.5043 Epoch 212/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3679 - mean_absolute_error: 0.5041 Epoch 213/250 64/64 [==============================] - 2s 37ms/step - loss: -2.3691 - mean_absolute_error: 0.5040 Epoch 214/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3688 - mean_absolute_error: 0.5039 Epoch 215/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3704 - mean_absolute_error: 0.5038 Epoch 216/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3699 - mean_absolute_error: 0.5039 Epoch 217/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3698 - mean_absolute_error: 0.5037 Epoch 218/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3738 - mean_absolute_error: 0.5040 Epoch 219/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3729 - mean_absolute_error: 0.5039 Epoch 220/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3717 - mean_absolute_error: 0.5039 Epoch 221/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3700 - mean_absolute_error: 0.5038 Epoch 222/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3712 - mean_absolute_error: 0.5038 Epoch 223/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3736 - mean_absolute_error: 0.5038 Epoch 224/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3736 - mean_absolute_error: 0.5038 Epoch 225/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3712 - mean_absolute_error: 0.5037 Epoch 226/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3711 - mean_absolute_error: 0.5038 Epoch 227/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3728 - mean_absolute_error: 0.5038 Epoch 228/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3747 - mean_absolute_error: 0.5039 Epoch 229/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3753 - mean_absolute_error: 0.5038 Epoch 230/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3730 - mean_absolute_error: 0.5036 Epoch 231/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3715 - mean_absolute_error: 0.5038 Epoch 232/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3709 - mean_absolute_error: 0.5041 Epoch 233/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3738 - mean_absolute_error: 0.5039 Epoch 234/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3735 - mean_absolute_error: 0.5037 Epoch 235/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3722 - mean_absolute_error: 0.5036 Epoch 236/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3717 - mean_absolute_error: 0.5036 Epoch 237/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3744 - mean_absolute_error: 0.5036 Epoch 238/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3752 - mean_absolute_error: 0.5035 Epoch 239/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3760 - mean_absolute_error: 0.5034 Epoch 240/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3628 - mean_absolute_error: 0.5048 Epoch 241/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3714 - mean_absolute_error: 0.5046 Epoch 242/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3735 - mean_absolute_error: 0.5044 Epoch 243/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3763 - mean_absolute_error: 0.5039 Epoch 244/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3766 - mean_absolute_error: 0.5041 Epoch 245/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3771 - mean_absolute_error: 0.5038 Epoch 246/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3760 - mean_absolute_error: 0.5039 Epoch 247/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3791 - mean_absolute_error: 0.5037 Epoch 248/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3801 - mean_absolute_error: 0.5037 Epoch 249/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3793 - mean_absolute_error: 0.5037 Epoch 250/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3783 - mean_absolute_error: 0.5035 Training model 5 of 5. Epoch 1/250 64/64 [==============================] - 25s 64ms/step - loss: 187643.4062 - mean_absolute_error: 0.5753 Epoch 2/250 64/64 [==============================] - 2s 33ms/step - loss: 655350.3750 - mean_absolute_error: 0.4811 Epoch 3/250 64/64 [==============================] - 3s 50ms/step - loss: 15.5320 - mean_absolute_error: 0.4887 Epoch 4/250 64/64 [==============================] - 3s 40ms/step - loss: -1.8046 - mean_absolute_error: 0.4907 Epoch 5/250 64/64 [==============================] - 2s 35ms/step - loss: -1.8644 - mean_absolute_error: 0.4918 Epoch 6/250 64/64 [==============================] - 2s 34ms/step - loss: -1.9010 - mean_absolute_error: 0.4929 Epoch 7/250 64/64 [==============================] - 2s 34ms/step - loss: -1.9288 - mean_absolute_error: 0.4937 Epoch 8/250 64/64 [==============================] - 2s 33ms/step - loss: -1.9526 - mean_absolute_error: 0.4944 Epoch 9/250 64/64 [==============================] - 2s 34ms/step - loss: -1.9746 - mean_absolute_error: 0.4947 Epoch 10/250 64/64 [==============================] - 2s 34ms/step - loss: -1.9965 - mean_absolute_error: 0.4951 Epoch 11/250 64/64 [==============================] - 2s 33ms/step - loss: -2.0130 - mean_absolute_error: 0.4961 Epoch 12/250 64/64 [==============================] - 2s 35ms/step - loss: -2.0326 - mean_absolute_error: 0.4958 Epoch 13/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0516 - mean_absolute_error: 0.4963 Epoch 14/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0672 - mean_absolute_error: 0.4968 Epoch 15/250 64/64 [==============================] - 2s 33ms/step - loss: -2.0574 - mean_absolute_error: 0.4971 Epoch 16/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0590 - mean_absolute_error: 0.4974 Epoch 17/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0649 - mean_absolute_error: 0.4976 Epoch 18/250 64/64 [==============================] - 2s 34ms/step - loss: -2.0647 - mean_absolute_error: 0.4975 Epoch 19/250 64/64 [==============================] - 2s 35ms/step - loss: -2.0978 - mean_absolute_error: 0.4972 Epoch 20/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1308 - mean_absolute_error: 0.4980 Epoch 21/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1357 - mean_absolute_error: 0.4980 Epoch 22/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1422 - mean_absolute_error: 0.4985 Epoch 23/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1445 - mean_absolute_error: 0.4984 Epoch 24/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1541 - mean_absolute_error: 0.4985 Epoch 25/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1623 - mean_absolute_error: 0.4991 Epoch 26/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1665 - mean_absolute_error: 0.4990 Epoch 27/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1723 - mean_absolute_error: 0.4991 Epoch 28/250 64/64 [==============================] - 2s 36ms/step - loss: -2.1766 - mean_absolute_error: 0.4992 Epoch 29/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1814 - mean_absolute_error: 0.4994 Epoch 30/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1857 - mean_absolute_error: 0.4996 Epoch 31/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1891 - mean_absolute_error: 0.4995 Epoch 32/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1939 - mean_absolute_error: 0.4995 Epoch 33/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1978 - mean_absolute_error: 0.4998 Epoch 34/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1926 - mean_absolute_error: 0.4995 Epoch 35/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1875 - mean_absolute_error: 0.5000 Epoch 36/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2086 - mean_absolute_error: 0.4995 Epoch 37/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2089 - mean_absolute_error: 0.5000 Epoch 38/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1938 - mean_absolute_error: 0.5000 Epoch 39/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1917 - mean_absolute_error: 0.4997 Epoch 40/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1912 - mean_absolute_error: 0.4999 Epoch 41/250 64/64 [==============================] - 2s 35ms/step - loss: -2.1932 - mean_absolute_error: 0.4997 Epoch 42/250 64/64 [==============================] - 2s 33ms/step - loss: -2.1954 - mean_absolute_error: 0.4998 Epoch 43/250 64/64 [==============================] - 2s 34ms/step - loss: -2.1970 - mean_absolute_error: 0.4996 Epoch 44/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2010 - mean_absolute_error: 0.4998 Epoch 45/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2001 - mean_absolute_error: 0.4997 Epoch 46/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2034 - mean_absolute_error: 0.4996 Epoch 47/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2053 - mean_absolute_error: 0.4997 Epoch 48/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2075 - mean_absolute_error: 0.4998 Epoch 49/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2081 - mean_absolute_error: 0.4997 Epoch 50/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2097 - mean_absolute_error: 0.4996 Epoch 51/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2093 - mean_absolute_error: 0.4996 Epoch 52/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2124 - mean_absolute_error: 0.4997 Epoch 53/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2114 - mean_absolute_error: 0.4997 Epoch 54/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2141 - mean_absolute_error: 0.4999 Epoch 55/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2137 - mean_absolute_error: 0.4996 Epoch 56/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2260 - mean_absolute_error: 0.5001 Epoch 57/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2482 - mean_absolute_error: 0.4997 Epoch 58/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2522 - mean_absolute_error: 0.4999 Epoch 59/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2572 - mean_absolute_error: 0.5002 Epoch 60/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2588 - mean_absolute_error: 0.5004 Epoch 61/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2600 - mean_absolute_error: 0.5001 Epoch 62/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2641 - mean_absolute_error: 0.5004 Epoch 63/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2660 - mean_absolute_error: 0.5001 Epoch 64/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2677 - mean_absolute_error: 0.5003 Epoch 65/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2698 - mean_absolute_error: 0.5004 Epoch 66/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2708 - mean_absolute_error: 0.5000 Epoch 67/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2735 - mean_absolute_error: 0.4998 Epoch 68/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2745 - mean_absolute_error: 0.4997 Epoch 69/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2759 - mean_absolute_error: 0.4996 Epoch 70/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2784 - mean_absolute_error: 0.4996 Epoch 71/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2782 - mean_absolute_error: 0.4996 Epoch 72/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2799 - mean_absolute_error: 0.4995 Epoch 73/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2823 - mean_absolute_error: 0.4995 Epoch 74/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2825 - mean_absolute_error: 0.4996 Epoch 75/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2863 - mean_absolute_error: 0.4995 Epoch 76/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2852 - mean_absolute_error: 0.4995 Epoch 77/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2878 - mean_absolute_error: 0.4994 Epoch 78/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2887 - mean_absolute_error: 0.4996 Epoch 79/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2889 - mean_absolute_error: 0.4995 Epoch 80/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2913 - mean_absolute_error: 0.4996 Epoch 81/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2920 - mean_absolute_error: 0.4997 Epoch 82/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2932 - mean_absolute_error: 0.4997 Epoch 83/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2939 - mean_absolute_error: 0.4996 Epoch 84/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2947 - mean_absolute_error: 0.4997 Epoch 85/250 64/64 [==============================] - 2s 34ms/step - loss: -2.2943 - mean_absolute_error: 0.4999 Epoch 86/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2955 - mean_absolute_error: 0.4995 Epoch 87/250 64/64 [==============================] - 2s 35ms/step - loss: -2.2964 - mean_absolute_error: 0.4997 Epoch 88/250 64/64 [==============================] - 2s 33ms/step - loss: -2.2989 - mean_absolute_error: 0.4995 Epoch 89/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3002 - mean_absolute_error: 0.4998 Epoch 90/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3005 - mean_absolute_error: 0.4996 Epoch 91/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3012 - mean_absolute_error: 0.4996 Epoch 92/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3027 - mean_absolute_error: 0.4997 Epoch 93/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3022 - mean_absolute_error: 0.4996 Epoch 94/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3041 - mean_absolute_error: 0.4997 Epoch 95/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3032 - mean_absolute_error: 0.4996 Epoch 96/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3033 - mean_absolute_error: 0.4995 Epoch 97/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3055 - mean_absolute_error: 0.4996 Epoch 98/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3072 - mean_absolute_error: 0.4997 Epoch 99/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3057 - mean_absolute_error: 0.4995 Epoch 100/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3071 - mean_absolute_error: 0.4995 Epoch 101/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3092 - mean_absolute_error: 0.4995 Epoch 102/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3103 - mean_absolute_error: 0.4996 Epoch 103/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3104 - mean_absolute_error: 0.4995 Epoch 104/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3102 - mean_absolute_error: 0.4994 Epoch 105/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3124 - mean_absolute_error: 0.4996 Epoch 106/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3127 - mean_absolute_error: 0.4998 Epoch 107/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3128 - mean_absolute_error: 0.4995 Epoch 108/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3126 - mean_absolute_error: 0.4994 Epoch 109/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3141 - mean_absolute_error: 0.4995 Epoch 110/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3139 - mean_absolute_error: 0.4994 Epoch 111/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3139 - mean_absolute_error: 0.4996 Epoch 112/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3162 - mean_absolute_error: 0.4995 Epoch 113/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3172 - mean_absolute_error: 0.4995 Epoch 114/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3171 - mean_absolute_error: 0.4995 Epoch 115/250 64/64 [==============================] - 2s 36ms/step - loss: -2.3166 - mean_absolute_error: 0.4996 Epoch 116/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3197 - mean_absolute_error: 0.4995 Epoch 117/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3206 - mean_absolute_error: 0.4995 Epoch 118/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3202 - mean_absolute_error: 0.4996 Epoch 119/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3202 - mean_absolute_error: 0.4996 Epoch 120/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3213 - mean_absolute_error: 0.4998 Epoch 121/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3213 - mean_absolute_error: 0.4996 Epoch 122/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3204 - mean_absolute_error: 0.4996 Epoch 123/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3225 - mean_absolute_error: 0.4997 Epoch 124/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3236 - mean_absolute_error: 0.4995 Epoch 125/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3246 - mean_absolute_error: 0.4997 Epoch 126/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3252 - mean_absolute_error: 0.4996 Epoch 127/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3257 - mean_absolute_error: 0.4997 Epoch 128/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3259 - mean_absolute_error: 0.4998 Epoch 129/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3265 - mean_absolute_error: 0.4997 Epoch 130/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3270 - mean_absolute_error: 0.4998 Epoch 131/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3281 - mean_absolute_error: 0.4996 Epoch 132/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3269 - mean_absolute_error: 0.4997 Epoch 133/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3288 - mean_absolute_error: 0.4997 Epoch 134/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3313 - mean_absolute_error: 0.4998 Epoch 135/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3305 - mean_absolute_error: 0.4998 Epoch 136/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3310 - mean_absolute_error: 0.4998 Epoch 137/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3296 - mean_absolute_error: 0.4998 Epoch 138/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3318 - mean_absolute_error: 0.4998 Epoch 139/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3303 - mean_absolute_error: 0.4997 Epoch 140/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3328 - mean_absolute_error: 0.4998 Epoch 141/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3317 - mean_absolute_error: 0.4999 Epoch 142/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3335 - mean_absolute_error: 0.4999 Epoch 143/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3348 - mean_absolute_error: 0.4998 Epoch 144/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3363 - mean_absolute_error: 0.4999 Epoch 145/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3340 - mean_absolute_error: 0.5000 Epoch 146/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3349 - mean_absolute_error: 0.5000 Epoch 147/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3362 - mean_absolute_error: 0.4999 Epoch 148/250 64/64 [==============================] - 2s 37ms/step - loss: -2.3344 - mean_absolute_error: 0.5002 Epoch 149/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3375 - mean_absolute_error: 0.4998 Epoch 150/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3364 - mean_absolute_error: 0.4999 Epoch 151/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3374 - mean_absolute_error: 0.5000 Epoch 152/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3381 - mean_absolute_error: 0.5000 Epoch 153/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3392 - mean_absolute_error: 0.5001 Epoch 154/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3353 - mean_absolute_error: 0.5001 Epoch 155/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3385 - mean_absolute_error: 0.5002 Epoch 156/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3413 - mean_absolute_error: 0.5000 Epoch 157/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3406 - mean_absolute_error: 0.5000 Epoch 158/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3377 - mean_absolute_error: 0.5001 0s - loss: -2.3362 - mean_absolute_err Epoch 159/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3393 - mean_absolute_error: 0.5000 Epoch 160/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3406 - mean_absolute_error: 0.5000 Epoch 161/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3430 - mean_absolute_error: 0.5001 Epoch 162/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3432 - mean_absolute_error: 0.5001 Epoch 163/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3444 - mean_absolute_error: 0.5001 Epoch 164/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3442 - mean_absolute_error: 0.5002 Epoch 165/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3440 - mean_absolute_error: 0.5002 Epoch 166/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3439 - mean_absolute_error: 0.5004 Epoch 167/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3438 - mean_absolute_error: 0.5002 Epoch 168/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3464 - mean_absolute_error: 0.5003 Epoch 169/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3461 - mean_absolute_error: 0.5001 Epoch 170/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3463 - mean_absolute_error: 0.5002 Epoch 171/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3459 - mean_absolute_error: 0.5002 Epoch 172/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3476 - mean_absolute_error: 0.5000 Epoch 173/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3451 - mean_absolute_error: 0.5003 Epoch 174/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3452 - mean_absolute_error: 0.5000 Epoch 175/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3456 - mean_absolute_error: 0.5001 Epoch 176/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3479 - mean_absolute_error: 0.5000 Epoch 177/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3490 - mean_absolute_error: 0.5001 Epoch 178/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3462 - mean_absolute_error: 0.5002 Epoch 179/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3491 - mean_absolute_error: 0.5000 Epoch 180/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3501 - mean_absolute_error: 0.5000 Epoch 181/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3503 - mean_absolute_error: 0.5000 Epoch 182/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3517 - mean_absolute_error: 0.5000 Epoch 183/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3526 - mean_absolute_error: 0.5000 Epoch 184/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3508 - mean_absolute_error: 0.5001 Epoch 185/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3513 - mean_absolute_error: 0.5001 Epoch 186/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3517 - mean_absolute_error: 0.5000 Epoch 187/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3507 - mean_absolute_error: 0.5001 Epoch 188/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3499 - mean_absolute_error: 0.5000 Epoch 189/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3511 - mean_absolute_error: 0.5000 Epoch 190/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3520 - mean_absolute_error: 0.5001 Epoch 191/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3520 - mean_absolute_error: 0.5000 Epoch 192/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3529 - mean_absolute_error: 0.5000 Epoch 193/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3538 - mean_absolute_error: 0.4999 Epoch 194/250 64/64 [==============================] - 2s 37ms/step - loss: -2.3526 - mean_absolute_error: 0.5000 Epoch 195/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3537 - mean_absolute_error: 0.5000 Epoch 196/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3537 - mean_absolute_error: 0.4999 Epoch 197/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3551 - mean_absolute_error: 0.4998 Epoch 198/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3565 - mean_absolute_error: 0.4999 Epoch 199/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3568 - mean_absolute_error: 0.5000 Epoch 200/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3580 - mean_absolute_error: 0.4997 Epoch 201/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3595 - mean_absolute_error: 0.4997 Epoch 202/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3577 - mean_absolute_error: 0.4998 Epoch 203/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3596 - mean_absolute_error: 0.4997 Epoch 204/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3616 - mean_absolute_error: 0.4998 Epoch 205/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3597 - mean_absolute_error: 0.4998 Epoch 206/250 64/64 [==============================] - 2s 36ms/step - loss: -2.3613 - mean_absolute_error: 0.4996 Epoch 207/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3609 - mean_absolute_error: 0.4996 Epoch 208/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3619 - mean_absolute_error: 0.4995 Epoch 209/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3616 - mean_absolute_error: 0.4994 Epoch 210/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3620 - mean_absolute_error: 0.4997 Epoch 211/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3604 - mean_absolute_error: 0.4995 Epoch 212/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3598 - mean_absolute_error: 0.4995 Epoch 213/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3635 - mean_absolute_error: 0.4997 Epoch 214/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3634 - mean_absolute_error: 0.4997 Epoch 215/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3643 - mean_absolute_error: 0.4997 Epoch 216/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3643 - mean_absolute_error: 0.4999 Epoch 217/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3648 - mean_absolute_error: 0.4995 Epoch 218/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3593 - mean_absolute_error: 0.5002 Epoch 219/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3657 - mean_absolute_error: 0.4998 Epoch 220/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3670 - mean_absolute_error: 0.4998 Epoch 221/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3682 - mean_absolute_error: 0.4997 Epoch 222/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3691 - mean_absolute_error: 0.4997 Epoch 223/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3690 - mean_absolute_error: 0.4998 Epoch 224/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3693 - mean_absolute_error: 0.4996 Epoch 225/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3697 - mean_absolute_error: 0.4999 Epoch 226/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3693 - mean_absolute_error: 0.4997 Epoch 227/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3713 - mean_absolute_error: 0.4997 Epoch 228/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3709 - mean_absolute_error: 0.4998 Epoch 229/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3730 - mean_absolute_error: 0.4996 Epoch 230/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3709 - mean_absolute_error: 0.4997 Epoch 231/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3735 - mean_absolute_error: 0.4998 Epoch 232/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3733 - mean_absolute_error: 0.4997 Epoch 233/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3733 - mean_absolute_error: 0.4998 Epoch 234/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3759 - mean_absolute_error: 0.4995 Epoch 235/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3758 - mean_absolute_error: 0.4995 Epoch 236/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3753 - mean_absolute_error: 0.4996 Epoch 237/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3751 - mean_absolute_error: 0.4996 Epoch 238/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3746 - mean_absolute_error: 0.4997 Epoch 239/250 64/64 [==============================] - 2s 37ms/step - loss: -2.3739 - mean_absolute_error: 0.4995 Epoch 240/250 64/64 [==============================] - 2s 36ms/step - loss: -2.3761 - mean_absolute_error: 0.4998 Epoch 241/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3747 - mean_absolute_error: 0.4996 Epoch 242/250 64/64 [==============================] - 2s 35ms/step - loss: -2.3748 - mean_absolute_error: 0.4995 Epoch 243/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3759 - mean_absolute_error: 0.4995 Epoch 244/250 64/64 [==============================] - 2s 33ms/step - loss: -2.3746 - mean_absolute_error: 0.4995 Epoch 245/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3774 - mean_absolute_error: 0.4997 Epoch 246/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3777 - mean_absolute_error: 0.4996 Epoch 247/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3759 - mean_absolute_error: 0.4994 Epoch 248/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3770 - mean_absolute_error: 0.4995 Epoch 249/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3782 - mean_absolute_error: 0.4997 Epoch 250/250 64/64 [==============================] - 2s 34ms/step - loss: -2.3782 - mean_absolute_error: 0.4997 Process completed in 2970.69 seconds
Make some training diagnostic plots¶
In [12]:
Copied!
"""
Plot accuracy/loss versus epoch
"""
fig = pyl.figure(figsize=(10,3))
## skip the first n epochs for clarity of the figures
n=40
ax1 = pyl.subplot(121)
for i in range(cn_model.num_models):
ax1.plot(cn_model.classifiers[i].history['mean_absolute_error'][n:], color='darkslategray', linewidth=2)
ax1.set_title('Model Mean Absolute Error')
ax1.set_ylabel('Mean Absolute Error')
ax1.set_xlabel('Epoch')
ax2 = pyl.subplot(122)
for i in range(cn_model.num_models):
ax2.plot(cn_model.classifiers[i].history['loss'][n:], color='crimson', linewidth=2)
ax2.set_title('Model Loss')
ax2.set_ylabel('Loss')
ax2.set_xlabel('Epoch')
pyl.show()
pyl.close()
"""
Plot accuracy/loss versus epoch
"""
fig = pyl.figure(figsize=(10,3))
## skip the first n epochs for clarity of the figures
n=40
ax1 = pyl.subplot(121)
for i in range(cn_model.num_models):
ax1.plot(cn_model.classifiers[i].history['mean_absolute_error'][n:], color='darkslategray', linewidth=2)
ax1.set_title('Model Mean Absolute Error')
ax1.set_ylabel('Mean Absolute Error')
ax1.set_xlabel('Epoch')
ax2 = pyl.subplot(122)
for i in range(cn_model.num_models):
ax2.plot(cn_model.classifiers[i].history['loss'][n:], color='crimson', linewidth=2)
ax2.set_title('Model Loss')
ax2.set_ylabel('Loss')
ax2.set_xlabel('Epoch')
pyl.show()
pyl.close()
In [13]:
Copied!
## get predictions for test and training samples
preds_test = cn_model.predict(X_test)
preds_train = cn_model.predict(X_train)
## get predictions for test and training samples
preds_test = cn_model.predict(X_test)
preds_train = cn_model.predict(X_train)
429/429 [==============================] - 1s 2ms/step 429/429 [==============================] - 1s 2ms/step 429/429 [==============================] - 1s 2ms/step 429/429 [==============================] - 1s 2ms/step 429/429 [==============================] - 1s 2ms/step 8148/8148 [==============================] - 13s 2ms/step 8148/8148 [==============================] - 13s 2ms/step 8148/8148 [==============================] - 15s 2ms/step 8148/8148 [==============================] - 14s 2ms/step 8148/8148 [==============================] - 13s 2ms/step
In [14]:
Copied!
# Make some diagnostic plots of the network performance
# Make some diagnostic plots of the network performance
In [15]:
Copied!
## some diagnostic plots!
flux = preds_test[:,0]*Flux_std
bins = np.linspace(0, max(np.max(y_train), np.max(y_test))*Flux_std, 20)
vals = bins[:-1]*0.0
e_vals = np.zeros((len(bins)-1,2))
for i in range(len(bins)-1):
w = np.where((flux>bins[i])&(flux<bins[i+1]))
vals[i] = np.median(flux[w]/y_test[w]/Flux_std)
dm = 1.09*preds_test[w, 1]**0.5/preds_test[w, 0]
e_vals[i, 0] = np.mean(dm)
e_vals[i, 1] = np.std(dm)
vals = np.array(vals)
e_vals = np.array(e_vals)
## flux prediction vs. known values
fig = pyl.figure(figsize=(20,10))
sp = fig.add_subplot(111)
sp.set_title('Train/Test Flux Predictions', fontsize=22)
pyl.errorbar(y_train*Flux_std, preds_train[:,0]/y_train,
yerr = preds_train[:, 1]**0.5/y_train, linestyle = 'none', lw=2, marker='o', label = 'train')
pyl.errorbar(y_test*Flux_std, preds_test[:,0]/y_test,
yerr = preds_test[:, 1]**0.5/y_test, linestyle = 'none', lw=2, marker='o', label='test')
pyl.errorbar((bins[1:]+bins[:-1])/2.0, vals, yerr=e_vals[:,0], linestyle = 'none', lw=3, ms=15, marker='o',c='k', zorder=11)
xlim = sp.get_xlim()
pyl.plot(xlim, [1,1], 'k--',lw=2,zorder=10)
flux_x = np.linspace(0, bins[-1], 50)
snr_est_y = (flux_x*(flux_x+68*900)**-(0.5)) # estimate bg limited SNR, with bg=1300 ADU, and a radius = 1.4*0.7" aperture, or 90 pix
pyl.plot(flux_x, 1.+(1./snr_est_y), 'r--', lw=4, zorder=12)
sp.set_ylim(0.7, 2.0)
pyl.plot(np.array([0.35,0.35])*Flux_std, sp.get_ylim(), zorder=10)
pyl.plot(np.array([0.25,0.25])*Flux_std, sp.get_ylim(), zorder=10, alpha=0.5)
sp.set_xlim(0, xlim[1])
sp.grid(linestyle=':')
pyl.xlabel('Instrumental Flux (ADU)', fontsize=22)
pyl.ylabel('Predict Flux/Plant Flux', fontsize=22)
pyl.xticks(fontsize=22)
pyl.yticks(fontsize=22)
pyl.legend(fontsize=22)
#pyl.savefig('ChapterFluxRegress.png', bbox_inches='tight')
## predicted-planted instrumental magnitude. Haven't spent as much time refining this figure.
fig = pyl.figure('2', figsize=(20,10))
sp = fig.add_subplot(111)
pyl.scatter(y_train*Flux_std, 1.09*preds_train[:,1]/preds_train[:, 0], label = 'train')
pyl.scatter(y_test*Flux_std, 1.09*preds_test[:,1]/preds_test[:, 0], label='test', alpha=0.5)
pyl.errorbar((bins[1:]+bins[:-1])/2.0, e_vals[:, 0], yerr = e_vals[:, 1],
linestyle = 'none', lw=2, marker='o',c='k')
sp.set_ylim(-0.5, 0.5)
pyl.xlabel('Instrumental Flux')
pyl.ylabel('Predict - Plant magnitude')
pyl.legend()
pyl.show()
## some diagnostic plots!
flux = preds_test[:,0]*Flux_std
bins = np.linspace(0, max(np.max(y_train), np.max(y_test))*Flux_std, 20)
vals = bins[:-1]*0.0
e_vals = np.zeros((len(bins)-1,2))
for i in range(len(bins)-1):
w = np.where((flux>bins[i])&(flux
/opt/conda/lib/python3.7/site-packages/ipykernel_launcher.py:38: RuntimeWarning: divide by zero encountered in true_divide
In [16]:
Copied!
if save_model:
cn_model.saveModel(Flux_std, mean, med, std,
save_model_path)
with open(f'{save_model_path}/training_data.pickle', 'bw+') as han:
pickle.dump([X_train, X_test, y_train, y_test, saved_par_dic], han)
if save_model:
cn_model.saveModel(Flux_std, mean, med, std,
save_model_path)
with open(f'{save_model_path}/training_data.pickle', 'bw+') as han:
pickle.dump([X_train, X_test, y_train, y_test, saved_par_dic], han)
In [17]:
Copied!
## plot of differences in predicted and planted apparent magnitude
train_mags = -2.5*np.log10(y_train*Flux_std)
test_mags = -2.5*np.log10(y_test*Flux_std)
pred_train_mags = -2.5*np.log10(preds_train[:, 0]*Flux_std)
pred_test_mags = -2.5*np.log10(preds_test[:, 0]*Flux_std)
w = np.where( ((kb_xys[test_index, -1]+Z_test)>22.5) & ((kb_xys[test_index, -1]+Z_test)<23))
figg = pyl.figure(figsize=(10, 10))
sp = figg.add_subplot(111)
pyl.scatter((kb_xys[train_index, -1]+Z_train), (pred_train_mags-kb_xys[train_index, -1]), alpha=0.2, label='train')
pyl.scatter(kb_xys[test_index, -1]+Z_test, pred_test_mags-kb_xys[test_index, -1], alpha=0.4, label='test')
pyl.legend()
pyl.plot([22,27], [0,0], 'k--', zorder=10)
pyl.xlabel('m_r planted')
pyl.ylabel('delta m')
sp.set_ylim(-0.5,0.5)
sp.set_yticks(np.arange(-0.5,0.5,0.025))
pyl.grid(linestyle=':')
## plot of differences in predicted and planted apparent magnitude
train_mags = -2.5*np.log10(y_train*Flux_std)
test_mags = -2.5*np.log10(y_test*Flux_std)
pred_train_mags = -2.5*np.log10(preds_train[:, 0]*Flux_std)
pred_test_mags = -2.5*np.log10(preds_test[:, 0]*Flux_std)
w = np.where( ((kb_xys[test_index, -1]+Z_test)>22.5) & ((kb_xys[test_index, -1]+Z_test)<23))
figg = pyl.figure(figsize=(10, 10))
sp = figg.add_subplot(111)
pyl.scatter((kb_xys[train_index, -1]+Z_train), (pred_train_mags-kb_xys[train_index, -1]), alpha=0.2, label='train')
pyl.scatter(kb_xys[test_index, -1]+Z_test, pred_test_mags-kb_xys[test_index, -1], alpha=0.4, label='test')
pyl.legend()
pyl.plot([22,27], [0,0], 'k--', zorder=10)
pyl.xlabel('m_r planted')
pyl.ylabel('delta m')
sp.set_ylim(-0.5,0.5)
sp.set_yticks(np.arange(-0.5,0.5,0.025))
pyl.grid(linestyle=':')
/opt/conda/lib/python3.7/site-packages/ipykernel_launcher.py:7: RuntimeWarning: invalid value encountered in log10 import sys /opt/conda/lib/python3.7/site-packages/ipykernel_launcher.py:8: RuntimeWarning: invalid value encountered in log10
Now read in validate visit 2022-08-24-AS1 and see how well the machine does¶
In [18]:
Copied!
visits_validate = ['2022-08-24-AS1']
chips = []
for c in np.arange(0, 40):#
chips.append(str(c).zfill(2))
stamp_files_validate, fs_validate, kb_xys_validate, zeropoints_validate = [], [], [], []
counter = 0
for j, v in enumerate(visits_validate):
for i, c in enumerate(chips):
stamps_path = f'{warps_path}/{v}'
plantLists_path = f'{warps_path}/{v}/'
kbmod_results_path = f'./kbmod_results/{v}/results_{c}/'
stamp_files_validate.append(f'{stamps_path}/stamps{gridType}_{c}_w_sr.pickle')
zpo = zpos[v][c]
### load the kbmod results
kb_xy = []
with open(f'{kbmod_results_path}/results_.txt') as han:
data = han.readlines()
for ii in range(len(data)):
s = data[ii].split()
x, y = float(s[5]), float(s[7])
repeat = False
for jj in range(len(kb_xy)):
if kb_xy[jj][0]==x and kb_xy[jj][1]==y:
repeat = True
break
if not repeat:
kb_xy.append([float(s[5]) , float(s[7]) , float(s[9]), float(s[11]), float(s[1]), 0.0])
kb_xy = np.array(kb_xy)
### load the plantlist sources
p_xy = []
plantLists = glob.glob(f'{plantLists_path}/{c}/*plantList')
plantLists.sort()
with open(plantLists[0]) as han:
data = han.readlines()
for ii in range(1,len(data)):
s = data[ii].split()
x,y = float(s[3]), float(s[4])
rate = float(s[5])*24./0.187
repeat = False
for jj in range(len(p_xy)):
if p_xy[jj][0] == x and p_xy[jj][1]==y:
p_xy[jj][2]-=0.75
repeat = True
if not repeat:
p_xy.append([x, y, float(s[9]), rate, 0])
if len(p_xy)>0:
p_xy = np.array(p_xy)
p_xy = p_xy[np.argsort(p_xy[:,2])]
p_xy = p_xy[np.where((p_xy[:,2]>nukeBright)&(p_xy[:,2]<nukeFaint))]
#label the good and bad sources
for ii in range(len(p_xy)):
d = ((p_xy[ii, 0] - kb_xy[:, 0])**2 + (p_xy[ii, 1] - kb_xy[:, 1])**2 )**0.5
d_rate = (p_xy[ii, 3] - (kb_xy[:, 2]**2 + kb_xy[:, 3]**2)**0.5)
ww = np.where((d<dist_lim)&(np.abs(d_rate)<rate_diff_lim))
if len(ww[0])>0:
for arg in ww[0]:
kb_xy[arg,-1] = p_xy[ii,2]-zpo
w_good = np.where(kb_xy[:,-1]!=0)
#load the stamps
with open(stamp_files_validate[-1], 'rb') as han:
f = pickle.load(han)
if rots[counter%len(chips)]!=0:# and rots[counter%len(chips)]!=2:
f = np.rot90(f, k=-rots[counter%len(chips)], axes=(1, 2))
counter+=1
### clip to avoid the crazy min pixel values
f = np.clip(f, -3500., np.max(f))
f_med = np.nanmedian(f)
f = f[w_good]
kb_xy = kb_xy[w_good]
fs_validate.append(f)
kb_xys_validate.append(kb_xy)
zeropoints_validate.append(np.zeros(len(f), dtype='float64')+zpo)
print(v, c,f.shape)
sns_frames_validate = np.concatenate(fs_validate)
kb_xys_validate = np.concatenate(kb_xys_validate)
zeropoints_validate = np.concatenate(zeropoints_validate)
del fs_validate
sns_frames_validate = sns_frames_validate.astype(image_data_type)
sns_labels_validate = 10.0**(-0.4*kb_xys_validate[:, -1])
sns_labels_validate = sns_labels_validate/Flux_std
normed_sns_frames_validate = sns_frames_validate
mean_validate,std_validate = np.nanmean(normed_sns_frames_validate), np.nanstd(normed_sns_frames_validate)
normed_sns_frames_validate -= mean_validate
normed_sns_frames_validate /= std
print('Normalized frame min and max:', np.nanmin(normed_sns_frames_validate), np.nanmax(normed_sns_frames_validate),'\n')
print(std,std_validate)
normed_sns_frames_validate = np.expand_dims(normed_sns_frames_validate, axis=-1)
preds_validate = cn_model.predict(normed_sns_frames_validate)
print(preds_validate)
fig = pyl.figure(figsize=(20,10))
sp = fig.add_subplot(111)
pyl.errorbar(sns_labels_validate*Flux_std, preds_validate[:, 0]/sns_labels_validate,
yerr = preds_validate[:, 1]**0.5/(sns_labels_validate),
linestyle = 'none', lw=2, marker='o', label = 'validate', alpha=0.5)
flux_x = np.linspace(0, np.max(sns_labels_validate*Flux_std), 100)
snr_est_y = (flux_x*(flux_x+68*900)**-(0.5)) # estimate bg limited SNR, with bg=1300 ADU, and a radius = 1.4*0.7" aperture, or 90 pix
pyl.plot(flux_x, 1.+(1./snr_est_y), 'r--', lw=4, zorder=12)
pyl.grid(linestyle=':')
xlim = sp.get_xlim()
pyl.plot([0,xlim[1]], [1., 1.], 'k--', lw=2, zorder=10)
sp.set_xlim(0.0, xlim[1])
sp.set_ylim(0.7, 2.0)
pyl.xlabel('Instrumental Flux (ADU)', fontsize=22)
pyl.ylabel('Predict Flux/Plant Flux', fontsize=22)
pyl.title('Validation Flux Predictions', fontsize=22)
pyl.xticks(fontsize=22)
pyl.yticks(fontsize=22)
#pyl.savefig('ChapterRegressValidate.pdf', bbox_inches='tight')
visits_validate = ['2022-08-24-AS1']
chips = []
for c in np.arange(0, 40):#
chips.append(str(c).zfill(2))
stamp_files_validate, fs_validate, kb_xys_validate, zeropoints_validate = [], [], [], []
counter = 0
for j, v in enumerate(visits_validate):
for i, c in enumerate(chips):
stamps_path = f'{warps_path}/{v}'
plantLists_path = f'{warps_path}/{v}/'
kbmod_results_path = f'./kbmod_results/{v}/results_{c}/'
stamp_files_validate.append(f'{stamps_path}/stamps{gridType}_{c}_w_sr.pickle')
zpo = zpos[v][c]
### load the kbmod results
kb_xy = []
with open(f'{kbmod_results_path}/results_.txt') as han:
data = han.readlines()
for ii in range(len(data)):
s = data[ii].split()
x, y = float(s[5]), float(s[7])
repeat = False
for jj in range(len(kb_xy)):
if kb_xy[jj][0]==x and kb_xy[jj][1]==y:
repeat = True
break
if not repeat:
kb_xy.append([float(s[5]) , float(s[7]) , float(s[9]), float(s[11]), float(s[1]), 0.0])
kb_xy = np.array(kb_xy)
### load the plantlist sources
p_xy = []
plantLists = glob.glob(f'{plantLists_path}/{c}/*plantList')
plantLists.sort()
with open(plantLists[0]) as han:
data = han.readlines()
for ii in range(1,len(data)):
s = data[ii].split()
x,y = float(s[3]), float(s[4])
rate = float(s[5])*24./0.187
repeat = False
for jj in range(len(p_xy)):
if p_xy[jj][0] == x and p_xy[jj][1]==y:
p_xy[jj][2]-=0.75
repeat = True
if not repeat:
p_xy.append([x, y, float(s[9]), rate, 0])
if len(p_xy)>0:
p_xy = np.array(p_xy)
p_xy = p_xy[np.argsort(p_xy[:,2])]
p_xy = p_xy[np.where((p_xy[:,2]>nukeBright)&(p_xy[:,2]0:
for arg in ww[0]:
kb_xy[arg,-1] = p_xy[ii,2]-zpo
w_good = np.where(kb_xy[:,-1]!=0)
#load the stamps
with open(stamp_files_validate[-1], 'rb') as han:
f = pickle.load(han)
if rots[counter%len(chips)]!=0:# and rots[counter%len(chips)]!=2:
f = np.rot90(f, k=-rots[counter%len(chips)], axes=(1, 2))
counter+=1
### clip to avoid the crazy min pixel values
f = np.clip(f, -3500., np.max(f))
f_med = np.nanmedian(f)
f = f[w_good]
kb_xy = kb_xy[w_good]
fs_validate.append(f)
kb_xys_validate.append(kb_xy)
zeropoints_validate.append(np.zeros(len(f), dtype='float64')+zpo)
print(v, c,f.shape)
sns_frames_validate = np.concatenate(fs_validate)
kb_xys_validate = np.concatenate(kb_xys_validate)
zeropoints_validate = np.concatenate(zeropoints_validate)
del fs_validate
sns_frames_validate = sns_frames_validate.astype(image_data_type)
sns_labels_validate = 10.0**(-0.4*kb_xys_validate[:, -1])
sns_labels_validate = sns_labels_validate/Flux_std
normed_sns_frames_validate = sns_frames_validate
mean_validate,std_validate = np.nanmean(normed_sns_frames_validate), np.nanstd(normed_sns_frames_validate)
normed_sns_frames_validate -= mean_validate
normed_sns_frames_validate /= std
print('Normalized frame min and max:', np.nanmin(normed_sns_frames_validate), np.nanmax(normed_sns_frames_validate),'\n')
print(std,std_validate)
normed_sns_frames_validate = np.expand_dims(normed_sns_frames_validate, axis=-1)
preds_validate = cn_model.predict(normed_sns_frames_validate)
print(preds_validate)
fig = pyl.figure(figsize=(20,10))
sp = fig.add_subplot(111)
pyl.errorbar(sns_labels_validate*Flux_std, preds_validate[:, 0]/sns_labels_validate,
yerr = preds_validate[:, 1]**0.5/(sns_labels_validate),
linestyle = 'none', lw=2, marker='o', label = 'validate', alpha=0.5)
flux_x = np.linspace(0, np.max(sns_labels_validate*Flux_std), 100)
snr_est_y = (flux_x*(flux_x+68*900)**-(0.5)) # estimate bg limited SNR, with bg=1300 ADU, and a radius = 1.4*0.7" aperture, or 90 pix
pyl.plot(flux_x, 1.+(1./snr_est_y), 'r--', lw=4, zorder=12)
pyl.grid(linestyle=':')
xlim = sp.get_xlim()
pyl.plot([0,xlim[1]], [1., 1.], 'k--', lw=2, zorder=10)
sp.set_xlim(0.0, xlim[1])
sp.set_ylim(0.7, 2.0)
pyl.xlabel('Instrumental Flux (ADU)', fontsize=22)
pyl.ylabel('Predict Flux/Plant Flux', fontsize=22)
pyl.title('Validation Flux Predictions', fontsize=22)
pyl.xticks(fontsize=22)
pyl.yticks(fontsize=22)
#pyl.savefig('ChapterRegressValidate.pdf', bbox_inches='tight')
2022-08-24-AS1 00 (61, 43, 43) 2022-08-24-AS1 01 (50, 43, 43) 2022-08-24-AS1 02 (54, 43, 43) 2022-08-24-AS1 03 (62, 43, 43) 2022-08-24-AS1 04 (50, 43, 43) 2022-08-24-AS1 05 (69, 43, 43) 2022-08-24-AS1 06 (63, 43, 43) 2022-08-24-AS1 07 (60, 43, 43) 2022-08-24-AS1 08 (60, 43, 43) 2022-08-24-AS1 09 (57, 43, 43) 2022-08-24-AS1 10 (55, 43, 43) 2022-08-24-AS1 11 (52, 43, 43) 2022-08-24-AS1 12 (68, 43, 43) 2022-08-24-AS1 13 (52, 43, 43) 2022-08-24-AS1 14 (68, 43, 43) 2022-08-24-AS1 15 (60, 43, 43) 2022-08-24-AS1 16 (62, 43, 43) 2022-08-24-AS1 17 (64, 43, 43) 2022-08-24-AS1 18 (56, 43, 43) 2022-08-24-AS1 19 (52, 43, 43) 2022-08-24-AS1 20 (60, 43, 43) 2022-08-24-AS1 21 (59, 43, 43) 2022-08-24-AS1 22 (55, 43, 43) 2022-08-24-AS1 23 (57, 43, 43) 2022-08-24-AS1 24 (50, 43, 43) 2022-08-24-AS1 25 (63, 43, 43) 2022-08-24-AS1 26 (55, 43, 43) 2022-08-24-AS1 27 (58, 43, 43) 2022-08-24-AS1 28 (56, 43, 43) 2022-08-24-AS1 29 (58, 43, 43) 2022-08-24-AS1 30 (59, 43, 43) 2022-08-24-AS1 31 (52, 43, 43) 2022-08-24-AS1 32 (50, 43, 43) 2022-08-24-AS1 33 (47, 43, 43) 2022-08-24-AS1 34 (50, 43, 43) 2022-08-24-AS1 35 (41, 43, 43) 2022-08-24-AS1 36 (59, 43, 43) 2022-08-24-AS1 37 (74, 43, 43) 2022-08-24-AS1 38 (67, 43, 43) 2022-08-24-AS1 39 (47, 43, 43) Normalized frame min and max: -13.106255 30.085432 14.035528 10.967248 72/72 [==============================] - 0s 3ms/step 72/72 [==============================] - 0s 2ms/step 72/72 [==============================] - 0s 2ms/step 72/72 [==============================] - 0s 2ms/step 72/72 [==============================] - 0s 2ms/step [[3.1897764e+00 3.6643110e-02] [3.0283675e+00 1.8730829e-02] [2.8002067e+00 1.0425494e-01] ... [1.0674439e-01 1.6833793e-03] [9.5082417e-02 1.3452958e-03] [1.8484716e-01 3.3594933e-03]]
/opt/conda/lib/python3.7/site-packages/ipykernel_launcher.py:128: RuntimeWarning: divide by zero encountered in true_divide
Out[18]:
(array([0.6, 0.8, 1. , 1.2, 1.4, 1.6, 1.8, 2. ]), [Text(0, 0, ''), Text(0, 0, ''), Text(0, 0, ''), Text(0, 0, ''), Text(0, 0, ''), Text(0, 0, ''), Text(0, 0, ''), Text(0, 0, '')])
In [ ]:
Copied!