Revision history [back]
 | 1 | initial version |
The last dimension of the input should be set to a fixed size, rather than None, in order to avoid a ValueError when using the Dense function in TensorFlow 2.2. This is because None signifies undefined, variable-sized dimensions, which are not compatible with the Dense layer that expects a fixed input shape. Therefore, it is important to specify a fixed size for the last dimension of the input as per the requirements of the Dense layer.