Some missed reset_classifier() type annotations

Author: rwightman

timm/models/efficientnet.py

@@ -156,7 +156,7 @@ def set_grad_checkpointing(self, enable=True):
     def get_classifier(self) -> nn.Module:
         return self.classifier
 
-    def reset_classifier(self, num_classes, global_pool='avg'):
+    def reset_classifier(self, num_classes: int, global_pool: str = 'avg'):
         self.num_classes = num_classes
         self.global_pool, self.classifier = create_classifier(
             self.num_features, self.num_classes, pool_type=global_pool)

timm/models/ghostnet.py

@@ -273,7 +273,7 @@ def set_grad_checkpointing(self, enable=True):
     def get_classifier(self) -> nn.Module:
         return self.classifier
 
-    def reset_classifier(self, num_classes, global_pool='avg'):
+    def reset_classifier(self, num_classes: int, global_pool: str = 'avg'):
         self.num_classes = num_classes
         # cannot meaningfully change pooling of efficient head after creation
         self.global_pool = SelectAdaptivePool2d(pool_type=global_pool)

timm/models/hrnet.py

@@ -739,7 +739,7 @@ def set_grad_checkpointing(self, enable=True):
     def get_classifier(self) -> nn.Module:
         return self.classifier
 
-    def reset_classifier(self, num_classes, global_pool='avg'):
+    def reset_classifier(self, num_classes: int, global_pool: str = 'avg'):
         self.num_classes = num_classes
         self.global_pool, self.classifier = create_classifier(
             self.num_features, self.num_classes, pool_type=global_pool)

timm/models/inception_v4.py

@@ -280,7 +280,7 @@ def set_grad_checkpointing(self, enable=True):
     def get_classifier(self) -> nn.Module:
         return self.last_linear
 
-    def reset_classifier(self, num_classes, global_pool='avg'):
+    def reset_classifier(self, num_classes: int, global_pool: str = 'avg'):
         self.num_classes = num_classes
         self.global_pool, self.last_linear = create_classifier(
             self.num_features, self.num_classes, pool_type=global_pool)

timm/models/metaformer.py

@@ -26,9 +26,9 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-
 from collections import OrderedDict
 from functools import partial
+from typing import Optional
 
 import torch
 import torch.nn as nn
@@ -548,7 +548,7 @@ def __init__(
         # if using MlpHead, dropout is handled by MlpHead
         if num_classes > 0:
             if self.use_mlp_head:
-                # FIXME hidden size
+                # FIXME not actually returning mlp hidden state right now as pre-logits.
                 final = MlpHead(self.num_features, num_classes, drop_rate=self.drop_rate)
                 self.head_hidden_size = self.num_features
             else:
@@ -583,7 +583,7 @@ def set_grad_checkpointing(self, enable=True):
     def get_classifier(self) -> nn.Module:
         return self.head.fc
 
-    def reset_classifier(self, num_classes=0, global_pool=None):
+    def reset_classifier(self, num_classes: int, global_pool: Optional[str] = None):
         if global_pool is not None:
             self.head.global_pool = SelectAdaptivePool2d(pool_type=global_pool)
             self.head.flatten = nn.Flatten(1) if global_pool else nn.Identity()

timm/models/nasnet.py

@@ -518,7 +518,7 @@ def set_grad_checkpointing(self, enable=True):
     def get_classifier(self) -> nn.Module:
         return self.last_linear
 
-    def reset_classifier(self, num_classes, global_pool='avg'):
+    def reset_classifier(self, num_classes: int, global_pool: str = 'avg'):
         self.num_classes = num_classes
         self.global_pool, self.last_linear = create_classifier(
             self.num_features, self.num_classes, pool_type=global_pool)

timm/models/pnasnet.py

@@ -307,7 +307,7 @@ def set_grad_checkpointing(self, enable=True):
     def get_classifier(self) -> nn.Module:
         return self.last_linear
 
-    def reset_classifier(self, num_classes, global_pool='avg'):
+    def reset_classifier(self, num_classes: int, global_pool: str = 'avg'):
         self.num_classes = num_classes
         self.global_pool, self.last_linear = create_classifier(
             self.num_features, self.num_classes, pool_type=global_pool)

timm/models/regnet.py

@@ -514,7 +514,7 @@ def set_grad_checkpointing(self, enable=True):
     def get_classifier(self) -> nn.Module:
         return self.head.fc
 
-    def reset_classifier(self, num_classes, global_pool='avg'):
+    def reset_classifier(self, num_classes: int, global_pool: Optional[str] = None):
         self.head.reset(num_classes, pool_type=global_pool)
 
     def forward_intermediates(

timm/models/rexnet.py

@@ -12,6 +12,7 @@
 
 from functools import partial
 from math import ceil
+from typing import Optional
 
 import torch
 import torch.nn as nn
@@ -229,7 +230,7 @@ def set_grad_checkpointing(self, enable=True):
     def get_classifier(self) -> nn.Module:
         return self.head.fc
 
-    def reset_classifier(self, num_classes, global_pool='avg'):
+    def reset_classifier(self, num_classes: int, global_pool: Optional[str] = None):
         self.num_classes = num_classes
         self.head.reset(num_classes, global_pool)
 

timm/models/selecsls.py

@@ -161,7 +161,7 @@ def set_grad_checkpointing(self, enable=True):
     def get_classifier(self) -> nn.Module:
         return self.fc
 
-    def reset_classifier(self, num_classes, global_pool='avg'):
+    def reset_classifier(self, num_classes: int, global_pool: str = 'avg'):
         self.num_classes = num_classes
         self.global_pool, self.fc = create_classifier(self.num_features, self.num_classes, pool_type=global_pool)