tt-pytorch
tt-pytorch copied to clipboard
initialize with pre-trained embeddings?
is there a way to initialize the TTEmbedding layer with a pre-trained embedding tensor?