本文整理汇总了Python中keras.backend.ctc_decode方法的典型用法代码示例。如果您正苦于以下问题：Python backend.ctc_decode方法的具体用法？Python backend.ctc_decode怎么用？Python backend.ctc_decode使用的例子？那么恭喜您, 这里精选的方法代码示例或许可以为您提供帮助。您也可以进一步了解该方法所在类keras.backend的用法示例。
在下文中一共展示了backend.ctc_decode方法的9个代码示例，这些例子默认根据受欢迎程度排序。您可以为喜欢或者感觉有用的代码点赞，您的评价将有助于系统推荐出更棒的Python代码示例。
示例1: ctc_complete_decoding_lambda_func
# 需要导入模块: from keras import backend [as 别名]
# 或者: from keras.backend import ctc_decode [as 别名]
def ctc_complete_decoding_lambda_func(args, **arguments):
"""
Complete CTC decoding using Keras (function K.ctc_decode)
:param args:
y_pred, input_length
:param arguments:
greedy, beam_width, top_paths
:return:
K.ctc_decode with dtype='float32'
"""
# import tensorflow as tf # Require for loading a model saved
y_pred, input_length = args
my_params = arguments
assert (K.backend() == 'tensorflow')
return K.cast(K.ctc_decode(y_pred, tf.squeeze(input_length), greedy=my_params['greedy'],
beam_width=my_params['beam_width'], top_paths=my_params['top_paths'])[0][0],
dtype='float32') 示例2: predict
# 需要导入模块: from keras import backend [as 别名]
# 或者: from keras.backend import ctc_decode [as 别名]
def predict(self, data_input, input_len):
"""
预测结果
:param data_input:
:param input_len:
:return: 返回语音识别后的拼音符号列表
"""
batch_size = 1
in_len = np.zeros((batch_size), dtype=np.int32)
in_len[0] = input_len
x_in = np.zeros((batch_size, 1600, self.AUDIO_FEATURE_LENGTH, 1), dtype=np.float)
for i in range(batch_size):
x_in[i, 0:len(data_input)] = data_input
with self.graph.as_default():
base_pred = self.base_model.predict(x=x_in)
base_pred = base_pred[:, :, :]
decoder = K.ctc_decode(base_pred, in_len, greedy=True, beam_width=100, top_paths=1)
result = K.get_value(decoder[0][0])[0]
return result 示例3: evaluate
# 需要导入模块: from keras import backend [as 别名]
# 或者: from keras.backend import ctc_decode [as 别名]
def evaluate(self):
correct_predictions = 0
correct_char_predictions = 0
x_val, y_val = self.val_generator[np.random.randint(0, int(self.val_generator.nb_samples / self.val_generator.batch_size))]
#x_val, y_val = next(self.val_generator)
y_pred = self.prediction_model.predict(x_val)
shape = y_pred[:, 2:, :].shape
ctc_decode = K.ctc_decode(y_pred[:, 2:, :], input_length=np.ones(shape[0])*shape[1])[0][0]
ctc_out = K.get_value(ctc_decode)[:, :self.label_len]
for i in range(self.val_generator.batch_size):
print(ctc_out[i])
result_str = ''.join([self.characters[c] for c in ctc_out[i]])
result_str = result_str.replace('-', '')
if result_str == y_val[i]:
correct_predictions += 1
print(result_str, y_val[i])
for c1, c2 in zip(result_str, y_val[i]):
if c1 == c2:
correct_char_predictions += 1
return correct_predictions / self.val_generator.batch_size, correct_char_predictions 示例4: predict_text
# 需要导入模块: from keras import backend [as 别名]
# 或者: from keras.backend import ctc_decode [as 别名]
def predict_text(model, img):
y_pred = model.predict(img[np.newaxis, :, :, :])
shape = y_pred[:, 2:, :].shape
ctc_decode = K.ctc_decode(y_pred[:, 2:, :], input_length=np.ones(shape[0])*shape[1])[0][0]
ctc_out = K.get_value(ctc_decode)[:, :cfg.label_len]
result_str = ''.join([cfg.characters[c] for c in ctc_out[0]])
result_str = result_str.replace('-', '')
return result_str 示例5: _decode
# 需要导入模块: from keras import backend [as 别名]
# 或者: from keras.backend import ctc_decode [as 别名]
def _decode(y_pred, input_length, greedy=True, beam_width=100, top_paths=1):
"""Decodes the output of a softmax.
Can use either greedy search (also known as best path)
or a constrained dictionary search.
# Arguments
y_pred: tensor `(samples, time_steps, num_categories)`
containing the prediction, or output of the softmax.
input_length: tensor `(samples, )` containing the sequence length for
each batch item in `y_pred`.
greedy: perform much faster best-path search if `true`.
This does not use a dictionary.
beam_width: if `greedy` is `false`: a beam search decoder will be used
with a beam of this width.
top_paths: if `greedy` is `false`,
how many of the most probable paths will be returned.
# Returns
Tuple:
List: if `greedy` is `true`, returns a list of one element that
contains the decoded sequence.
If `false`, returns the `top_paths` most probable
decoded sequences.
Important: blank labels are returned as `-1`.
Tensor `(top_paths, )` that contains
the log probability of each decoded sequence.
"""
decoded = K.ctc_decode(y_pred=y_pred, input_length=input_length,
greedy=greedy, beam_width=beam_width, top_paths=top_paths)
paths = [path.eval(session=K.get_session()) for path in decoded[0]]
logprobs = decoded[1].eval(session=K.get_session())
return (paths, logprobs) 示例6: predict
# 需要导入模块: from keras import backend [as 别名]
# 或者: from keras.backend import ctc_decode [as 别名]
def predict(self , data_input , input_len):
batch_size = 1
in_len = np.zeros((batch_size) , dtype=np.int32)
in_len[0] = input_len
x_in = np.zeros(shape=[batch_size , 2000 , self.FEATURE_LENGTH , 1] , dtype=np.float)
for i in range(batch_size):
x_in[i , 0 : len(data_input)] = data_input
base_pred = self.base_model.predict(x=x_in)
base_pred = base_pred[: , : , :]
r = K.ctc_decode(base_pred , in_len , greedy=True , beam_width=100 , top_paths=1)
r1 = K.get_value(r[0][0])
r1 = r1[0]
return r1 示例7: decode_predict_ctc
# 需要导入模块: from keras import backend [as 别名]
# 或者: from keras.backend import ctc_decode [as 别名]
def decode_predict_ctc(out, chars, top_paths=1):
results = []
beam_width = 5
if beam_width < top_paths:
beam_width = top_paths
for i in range(top_paths):
lables = K.get_value(
K.ctc_decode(
out, input_length=np.ones(out.shape[0]) * out.shape[1],
greedy=False, beam_width=beam_width, top_paths=top_paths
)[0][i]
)[0]
text = labels_to_text(chars, lables)
results.append(text)
return results 示例8: Predict
# 需要导入模块: from keras import backend [as 别名]
# 或者: from keras.backend import ctc_decode [as 别名]
def Predict(self, data_input, input_len):
'''
预测结果
返回语音识别后的拼音符号列表
'''
batch_size = 1
in_len = np.zeros((batch_size),dtype = np.int32)
in_len[0] = input_len
x_in = np.zeros((batch_size, 1600, self.AUDIO_FEATURE_LENGTH, 1), dtype=np.float)
for i in range(batch_size):
x_in[i,0:len(data_input)] = data_input
base_pred = self.base_model.predict(x = x_in)
#print('base_pred:\n', base_pred)
#y_p = base_pred
#for j in range(200):
# mean = np.sum(y_p[0][j]) / y_p[0][j].shape[0]
# print('max y_p:',np.max(y_p[0][j]),'min y_p:',np.min(y_p[0][j]),'mean y_p:',mean,'mid y_p:',y_p[0][j][100])
# print('argmin:',np.argmin(y_p[0][j]),'argmax:',np.argmax(y_p[0][j]))
# count=0
# for i in range(y_p[0][j].shape[0]):
# if(y_p[0][j][i] < mean):
# count += 1
# print('count:',count)
base_pred =base_pred[:, :, :]
#base_pred =base_pred[:, 2:, :]
r = K.ctc_decode(base_pred, in_len, greedy = True, beam_width=100, top_paths=1)
#print('r', r)
r1 = r[0][0].eval(session=tf.compat.v1.Session())
tf.compat.v1.reset_default_graph()
return r1[0] 示例9: Predict
# 需要导入模块: from keras import backend [as 别名]
# 或者: from keras.backend import ctc_decode [as 别名]
def Predict(self, data_input, input_len):
'''
预测结果
返回语音识别后的拼音符号列表
'''
batch_size = 1
in_len = np.zeros((batch_size),dtype = np.int32)
in_len[0] = input_len
x_in = np.zeros((batch_size, 1600, self.AUDIO_FEATURE_LENGTH, 1), dtype=np.float)
for i in range(batch_size):
x_in[i,0:len(data_input)] = data_input
base_pred = self.base_model.predict(x = x_in)
#print('base_pred:\n', base_pred)
#y_p = base_pred
#for j in range(200):
# mean = np.sum(y_p[0][j]) / y_p[0][j].shape[0]
# print('max y_p:',np.max(y_p[0][j]),'min y_p:',np.min(y_p[0][j]),'mean y_p:',mean,'mid y_p:',y_p[0][j][100])
# print('argmin:',np.argmin(y_p[0][j]),'argmax:',np.argmax(y_p[0][j]))
# count=0
# for i in range(y_p[0][j].shape[0]):
# if(y_p[0][j][i] < mean):
# count += 1
# print('count:',count)
base_pred =base_pred[:, :, :]
#base_pred =base_pred[:, 2:, :]
r = K.ctc_decode(base_pred, in_len, greedy = True, beam_width=100, top_paths=1)
#print('r', r)
r1 = K.get_value(r[0][0])
#print('r1', r1)
#r2 = K.get_value(r[1])
#print(r2)
r1=r1[0]
return r1
pass