-
Notifications
You must be signed in to change notification settings - Fork 0
Open
Description
-
Look for: Types of inputs (facemesh data, eye data, etc.). Use template in "Reference guide for embedding model building (bad)" Collab notebook.
-
Try to construct attention model from embedding model. ex: embedding_model.save() - Saves whole model to be used as layer in another model. Definitely use convolution for images.
-
Edit reference guide and annotate to make ideas more clear.
-
Look into fixing issue of unusable last frames in webm's. Go back to downloading webm's?
Metadata
Metadata
Assignees
Labels
No labels