我在使用tensorflow-gpu时,运行如下的代码:
import tensorflow as tf
a = tf.constant([1.0,2.0,3.0],shape = [3],name='a')
print(a)
能正确显示tf.Tensor([1. 2. 3.], shape=(3,), dtype=float32)
但当我加入以下两行代码
b = tf.constant([1.0,2.0,3.0], shape = [3],name='b')
a+b
会报如下的错误
F tensorflow/stream_executor/cuda/cuda_driver.cc:209] Check failed: err == cudaSuccess || err == cudaErrorInvalidValue Unexpected CUDA error: invalid argument
请大神帮忙解答
================================================
参照这篇博客,重新安装了一下,问题已经解决了
https://blog.csdn.net/weixin_43411129/article/details/101390407