ValueError: Argument must be a dense tensor:... got shape [6, 60, 160, 3], but wanted [6].

在將 列表或元組 數據轉換成 dataset類型時

 

import numpy as np
import tensorflow as tf
from sklearn.cross_validation import train_test_split

 

pic_array=np.ones((60,160,3)) #圖片的長寬爲60*160，每一個像素點的由rgb3個值表示像素
pic_txt_array=np.ones((26,4)) #表示單個字母的向量長爲26，共4個字母
data_x=[pic_array  for i in range(1000)] #1000張圖片的集合
data_y=[pic_txt_array for i in range(1000)]#1000長圖片對應的字母的集合

#將裝着樣本的列表 轉換成dataset格式
train_dataset=tf.data.Dataset.from_tensor_slices((data_x,data_y))

 

發生異常：

File "tf_test.py", line 10, in <module>
train_dataset=tf.data.Dataset.from_tensor_slices((data_x,data_y))
File "F:\Program Files\Python35\lib\site-packages\tensorflow\python\data\ops\dataset_ops.py", line 235, in from_tensor_slices
return TensorSliceDataset(tensors)
File "F:\Program Files\Python35\lib\site-packages\tensorflow\python\data\ops\dataset_ops.py", line 1030, in __init__
for i, t in enumerate(nest.flatten(tensors))
........
File "F:\Program Files\Python35\lib\site-packages\tensorflow\python\framework\constant_op.py", line 214, in constant
value, dtype=dtype, shape=shape, verify_shape=verify_shape))
File "F:\Program Files\Python35\lib\site-packages\tensorflow\python\framework\tensor_util.py", line 441, in make_tensor_proto
_GetDenseDimensions(values)))
ValueError: Argument must be a dense tensor: [array([[[1., 1., 1.],...

...,
[1., 1., 1.],
[1., 1., 1.],
[1., 1., 1.]]])] - got shape [6, 60, 160, 3], but wanted [6].

 

解決：修改源數據的格式

 

data_x=np.asarray([pic_array for i in range(1000)]) #1000張圖片的集合
data_y=np.asarray([pic_txt_array for i in range(1000)]) #1000長圖片對應的字母的集合