cascaded norm

#11

by J22 - opened 12 days ago

There is a LayerNorm (post_layernorm) following the last layer of ViT, which is followed by a RMSNorm (ln_q from VLPatchMerger).

Is there any special consideration on cascading two Norms?

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment