A technique to reduce the precision of model weights, leading to smaller and faster models.
Latent Space
