ValueError：Tensor转换请求具有dtype float32的Tensor的dtype int32

Question

ValueError：张量转换请求dtype float32的张量为dtype int32：'Tensor（“ embedding_1 / random_uniform：0”，shape =（4518，1），dtype = float32）'

File "C:\PythonProjects\combined\venv\lib\site-packages\tensorflow\python\framework\ops.py", line 1050, in convert_to_tensor as_ref=False):

@tf_export("convert_to_tensor")
def convert_to_tensor(value, dtype=None, name=None, preferred_dtype=None):

File "C:\PythonProjects\combined\venv\lib\site-packages\tensorflow\python\framework\ops.py", line 1146, in internal_convert_to_tensor
ret = conversion_func(value, dtype=dtype, name=name, as_ref=as_ref)

internal_convert_to_tensor(value=value,dtype=dtype,name=name,preferred_dtype=preferred_dtype,as_ref=False)

我的问题是如何将dtype更改为int32？

Answer 1

您可以使用tf.cast(given_tensor, tf.int32)或tf.to_int32(given_tensor)将张量从float32投射到int32。