Tensorflow float32 dtypes

[hudson-ai]

My apologies if this feature exists somewhere, but I can't find it.

Is there a way to use cvxpy layers with tf.float32 parameters? I understand if the actual optimization step has to use float64 precison, but is there a way to accept float32 inputs/outputs/gradients? If not, is this a feature that could be added?

Thanks!

[hudson-ai]

It seems that these layers CAN accept float32 input, but their outputs are always float64 (as are computed gradients)