def inputProcedData(self):
filesize = int(self.frmSizeBox.text())
print "Input proced data", filesize
self.fname = [str(self.txtSepFile.text())]
input_data = data_proc2.load_proced_data_flag(self.fname,
datalen=filesize)
self.loadInputData(input_data)
dim_h = args.n_units
batchsize = args.batchsize
bprop_len = args.bprop_len
grad_clip = 1
train_file = glob.glob(args.train_dir + "/*")
s_train_file = glob.glob(args.s_train_dir + "/*")
valid_file = glob.glob(args.valid_dir + "/*")
test_file = glob.glob(args.test_dir + "/*")
if len(train_file) == 0:
print "error train folder no file!"
train_data = data_proc2.load_proced_data(train_file) #(joints, speaks, annos)
s_train_data = data_proc2.load_proced_data_flag(
s_train_file) #(joints, speaks, annos)
valid_data = data_proc2.load_proced_data(valid_file, datalen=args.datalen)
test_data = data_proc2.load_proced_data(test_file, datalen=args.datalen)
print("train_data:", train_data[0].shape, train_data[2].shape,
train_data[2].shape)
# Data set
train_joints, train_speaks, train_annos = train_data[0], train_data[
1], train_data[2]
s_train_joints, s_train_speaks, s_train_annos, s_train_flags = s_train_data[
"joints"], s_train_data["speaks"], s_train_data["annos"], s_train_data[
"flags"]
valid_joints, valid_speaks, valid_annos = valid_data[0], valid_data[
start = t
if data[t] == 1 and data[t+1] == 0:
sum_time += sum(diff_time[start:t])
return sum_time
fnames = sorted(glob.glob(args.filename+"/*"))
cap = ["person", "robot"]
types = ["model", "control", "random", "person"]
for n, fname in enumerate(fnames):
input_data = data_proc2.load_proced_data_flag([fname], datalen=-1) #(1000, 2)
a_data = np.abs(input_data["annos"])#[:1000]
t_data = input_data["times"]#[:1000]
diff_time = []
for i in range(len(t_data)-1):
if i == 0:
diff_time.append(0)
else:
diff_time.append(t_data[i+1]-t_data[i])
plt.subplot(4,1,n+1)
if n < 3:
barh_plot(a_data[:,0], a_data[:,1], t_data)
