Search papers, labs, and topics across Lattice.
1
0
3
Ditch the training and still shrink your image embeddings by 75% while boosting accuracy, thanks to a clever CCA trick that distills knowledge between pre-trained models.