-
Notifications
You must be signed in to change notification settings - Fork 2
Open
Description
模型部分基本一样,都是251
但我想用fit的方式进行训练,却得到如下错误
AssertionError: Could not compute output Tensor("ctc/ExpandDims:0", shape=(None, 1), dtype=float32)
不能计算输出张量,我怀疑是我的输入参数有问题,搜到你的关于backend.ctc_batch_cost的一篇文章才转到这里来
#我是用MFCC做特征的
x_train.shape (8,800,26,1)
#最长21个字
y_train.shape (21,1)
history = self.model.fit(x_train,
y_train,
epochs=epochs,
verbose=1,
validation_split=0.1,
callbacks=self.callbacks)
我想问一下,这个报错是和y_train还有ctc_batch_cost(y_true, y_pred, input_length, label_length)有关系吗?
需要如何输入数据呢?
Reactions are currently unavailable
Metadata
Metadata
Assignees
Labels
No labels