allow for no final output head on the vit

Author: lucidrains

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "vit-pytorch"
-version = "1.17.1"
+version = "1.17.2"
 description = "Vision Transformer (ViT) - Pytorch"
 readme = { file = "README.md", content-type = "text/markdown" }
 license = { file = "LICENSE" }

vit_pytorch/vit.py

@@ -113,7 +113,7 @@ def __init__(self, *, image_size, patch_size, num_classes, dim, depth, heads, ml
         self.pool = pool
         self.to_latent = nn.Identity()
 
-        self.mlp_head = nn.Linear(dim, num_classes)
+        self.mlp_head = nn.Linear(dim, num_classes) if num_classes > 0 else None
 
     def forward(self, img):
         batch = img.shape[0]
@@ -129,6 +129,9 @@ def forward(self, img):
 
         x = self.transformer(x)
 
+        if not exists(self.mlp_head):
+            return x
+
         x = x.mean(dim = 1) if self.pool == 'mean' else x[:, 0]
 
         x = self.to_latent(x)