import numpy as np
from keras.models import model_from_json, model_from_yaml
from mir_eval import melody
TIMESTEP = 128
SUBDIVISION = 8
def freq2midi(f):
return 69 + 12*np.log2(f/440)
def midi2freq(m):
return 2**((m - 69)/ 12) * 440
def sigmoid(x):
return 1 / (1 + np.exp(-x))
def note_res_downsampling(score):
# filter
f = [0.1, 0.2, 0.4, 0.2, 0.1]
r = len(f) // 2
new_score = np.zeros((score.shape[0], 88))
pad = np.zeros((new_score.shape[0], 2))
score = np.concatenate([pad, score], axis=1)
f_aug = np.tile(f, (new_score.shape[0], 1))
for i in range(0, 352, 4):
cent = i + 2
lb = max(0, cent - r)
ub = min(353, (cent + 1) + r)
new_score[:, i // 4] = np.sum(score[:, lb:ub] * f_aug, axis=1)
return new_score
def padding(x,
feature_num,
timesteps,
dimension=False):
extended_chorale = np.array(x)
if (((feature_num - x.shape[1]) % 2) == 0):
p_t = (feature_num - x.shape[1]) // 2
p_b = p_t
else:
p_t = (feature_num - x.shape[1]) // 2
p_b = p_t + 1
top = np.zeros((extended_chorale.shape[0], p_t))
bottom = np.zeros((extended_chorale.shape[0], p_b))
extended_chorale = np.concatenate([top, extended_chorale, bottom], axis=1)
padding_dimensions = (timesteps,) + extended_chorale.shape[1:]
padding_start = np.zeros(padding_dimensions)
padding_end = np.zeros(padding_dimensions)
padding_start[:, :p_t] = 1
padding_end[:, -p_b:] = 1
extended_chorale = np.concatenate((padding_start,
extended_chorale,
padding_end),
axis=0)
if (dimension):
return extended_chorale, p_t, p_b
else:
return extended_chorale
def matrix_parser(m):
x = np.zeros(shape=(m.shape[0], 2))
for i in range(len(m)):
if (np.sum(m[i]) != 0):
x[i][0] = 1
x[i][1] = midi2freq(np.argmax(m[i]) / 4 + 21)
x[:, 1] = melody.hz2cents(x[:, 1])
return x
def load_model(model_name):
"""
"""
ext = '.yaml'
model = model_from_yaml(open(model_name + ext).read())
model.load_weights(model_name + '_weights.h5')
print("model " + model_name + " loaded")
return model
def save_model(model, model_name, overwrite=False):
# SAVE MODEL
string = model.to_yaml()
ext = '.yaml'
open(model_name + ext, 'w').write(string)
model.save_weights(model_name + '_weights.h5', overwrite=overwrite)
print("model " + model_name + " saved")
def model_copy(origin, target):
for index, layer in enumerate(target.layers):
if layer.__class__.__name__ == 'LSTM':
weights = origin.layers[index].get_weights()
units = weights[1].shape[0]
bias = weights[2]
if len(bias) == units * 8:
# reshape the kernels
kernels = np.split(weights[0], 4, axis=1)
kernels = [kernel.reshape(-1).reshape(kernel.shape, order='F') for kernel in kernels]
weights[0] = np.concatenate(kernels, axis=1)
# transpose the recurrent kernels
recurrent_kernels = np.split(weights[1], 4, axis=1)
recurrent_kernels = [kernel.T for kernel in recurrent_kernels]
weights[1] = np.concatenate(recurrent_kernels, axis=1)
# split the bias into half and merge
weights[2] = bias[:units * 4] + bias[units * 4:]
layer.set_weights(weights)
print("Set success")
else:
layer.set_weights(origin.layers[index].get_weights())
