[英]Tensorflow 2: shape mismatch when serialize and decode it back
I have a tensor A of shape (300,256,256).我有一个形状为 (300,256,256) 的张量 A。 I want to serialize A to save as tfrecord format.
我想序列化 A 以保存为 tfrecord 格式。 But I cannot convert it back to tensor with same shape.
但我无法将其转换回具有相同形状的张量。
A = tf.convert_to_tensor( *a numpy array with float32 type* )
B = tf.io.serialize_tensor(A)
C = tf.reshape(tf.io.decode_raw(B, out_type=tf.float32),[300,256,256])
If I run the code above, I got a shape error:如果我运行上面的代码,我会得到一个形状错误:
tensorflow.python.framework.errors_impl.InvalidArgumentError: Input to reshape is a tensor with 19660806 values, but the requested shape has 19660800 [Op:Reshape]
tensorflow.python.framework.errors_impl.InvalidArgumentError:reshape 的输入是一个具有 19660806 值的张量,但请求的形状有 19660800 [Op:Reshape]
It seems that when I serialize or when I decode, 6 floats are added.似乎当我序列化或解码时,添加了 6 个浮点数。 (very weird)
(很奇怪)
Try using: tf.io.parse_tensor()
, instead of tf.io.decode_raw()
.尝试使用:
tf.io.parse_tensor()
,而不是tf.io.decode_raw()
。
https://www.tensorflow.org/api_docs/python/tf/io/parse_tensor https://www.tensorflow.org/api_docs/python/tf/io/parse_tensor
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.