TypeError: Cannot convert 1.0 to EagerTensor of dtype int64

最新推荐文章于 2022-03-18 14:24:19 发布

shelleyHLX

最新推荐文章于 2022-03-18 14:24:19 发布

阅读量7.6k

点赞数

分类专栏： tensorflow 文章标签： tensorflow

本文链接：https://blog.csdn.net/qq_27009517/article/details/109164777

版权

在使用TensorFlow 2.2.0时遇到错误：TypeError: Cannot convert 0.0 to EagerTensor of dtype int32。问题根源在于调用GradientTape.gradient时源张量的dtype必须为浮点类型，但实际得到的是tf.int32。解决方案是将int32类型的变量转换为float32。参考链接提供了解决此问题的具体例子。

摘要由CSDN通过智能技术生成

tensorflow2.2.0的错误：

WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int32
2020-10-19 17:18:38.582782: W tensorflow/core/framework/op_kernel.cc:1741] Invalid argument: TypeError: Cannot convert 0.0 to EagerTensor of dtype int32
Traceback (most recent call last):

File "/home/hlx/anaconda3/lib/python3.6/site-packages/tensorflow/python/ops/script_ops.py", line 241, in __call__
return func(device, token, args)

File "/home/hlx/anaconda3/lib/python3.6/site-packages/tensorflow/python/ops/script_ops.py", line 142, in __call__
for (x, dtype) in zip(ret, self._out_dtypes)

File "/home/hlx/anaconda3/lib/python3.6/site-packages/tensorflow/python/ops/script_ops.py", line 142, in <listcomp>
for (x, dtype) in zip(ret, self._out_dtypes)

File "/home/hlx/anaconda3/lib/python3.6/site-packages/tensorflow/python/ops/script_ops.py", line 118, in _convert
return constant_op.constant(0.0, dtype=dtype)

File "/home/hlx/anaconda3/lib/python3.6/site-packages/tensorflow/python/framework/constant_op.py", line 262, in constant
allow_broadcast=True)

File "/home/hlx/anaconda3/lib/python3.6/site-packages/tensorflow/python/framework/constant_op.py", line 270, in _constant_impl
t = convert_to_eager_tensor(value, ctx, dtype)

File "/home/hlx/anaconda3/lib/python3.6/site-packages/tensorflow/python/framework/constant_op.py", line 96, in convert_to_eager_tensor
return ops.EagerTensor(value, ctx.device_name, dtype)

TypeError: Cannot convert 0.0 to EagerTensor of dtype int32

原因：

The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.