如何创建 0 维 tf.Variable？

Question

dimension tf.Variable. Before I did that via:

tensor = tf.zeros([0,], tf.float32) * (rank + 1)
w = tf.Variable(tensor.numpy().tolist())

however it says that AttributeError: 'Tensor' object has no attribute 'numpy' when I put these two lines in a new project. I think it is because eager_mode is disabled in that project i am working on? Is there any way I can create the same Viriable directly without numpy() ?

Answer 1

To create a tensor of dimension 0, run the following code

## rank 0
# Default name
rank_0_tensor = tf.constant(4, tf.int16) 
print(rank_0_tensor)
        
Output: tf.Tensor(4, shape=(), dtype=int16)

Reference: https://www.tensorflow.org/guide/tensor