jy0205/Pyramid-Flow

Null token embedding missing when using pre-computed text embeddings

Closed this issue · 1 comments

Hi all,

when training with pre-computed text, we need the null token embedding self.null_text_embeds here.

Where do we get this from?

This is the pre-extracted textual embeddings of empty str, e.g., "", which is used for training with classifier-free guidance.