(patch_size=16, num_register_tokens=0, **kwargs)
| 351 | |
| 352 | |
| 353 | def vit_base(patch_size=16, num_register_tokens=0, **kwargs): |
| 354 | model = DinoVisionTransformer( |
| 355 | patch_size=patch_size, |
| 356 | embed_dim=768, |
| 357 | depth=12, |
| 358 | num_heads=12, |
| 359 | mlp_ratio=4, |
| 360 | block_fn=partial(Block, attn_class=MemEffAttention), |
| 361 | num_register_tokens=num_register_tokens, |
| 362 | **kwargs, |
| 363 | ) |
| 364 | return model |
| 365 | |
| 366 | |
| 367 | def vit_large(patch_size=16, num_register_tokens=0, **kwargs): |
nothing calls this directly
no test coverage detected