Pretrained weights for lucid.models.clip_vit_large_14_336.
Notes
Not interchangeable with CLIPViTLarge14Weights despite the
identical tower shape: the positional table has 577 rows here
against 257 there, so loading either into the other's model fails on
that one tensor.
Examples
>>> from lucid.models.multimodal.clip._weights import CLIPViTLarge14_336Weights
>>> list(CLIPViTLarge14_336Weights.__members__)
['OPENAI_WIT400M', 'DEFAULT']
DEFAULT is an alias rather than a fourth entry, so a bare
pretrained=True and the tag it resolves to cannot drift apart.
>>> CLIPViTLarge14_336Weights.DEFAULT is CLIPViTLarge14_336Weights.OPENAI_WIT400M
True
>>> CLIPViTLarge14_336Weights.OPENAI_WIT400M.num_classes
0