Cleanup weight init for byob/byoanet and related

Author: rwightman

timm/models/byobnet.py

@@ -33,7 +33,7 @@
 import torch.nn as nn
 
 from timm.data import IMAGENET_DEFAULT_MEAN, IMAGENET_DEFAULT_STD
-from .helpers import build_model_with_cfg
+from .helpers import build_model_with_cfg, named_apply
 from .layers import ClassifierHead, ConvBnAct, BatchNormAct2d, DropPath, AvgPool2dSame, \
     create_conv2d, get_act_layer, convert_norm_act, get_attn, make_divisible, to_2tuple
 from .registry import register_model
@@ -166,7 +166,7 @@ class ByoModelCfg:
     stem_chs: int = 32
     width_factor: float = 1.0
     num_features: int = 0  # num out_channels for final conv, no final 1x1 conv if 0
-    zero_init_last_bn: bool = True
+    zero_init_last: bool = True  # zero init last weight (usually bn) in residual path
     fixed_input_size: bool = False  # model constrained to a fixed-input size / img_size must be provided on creation
 
     act_layer: str = 'relu'
@@ -757,8 +757,8 @@ def __init__(
         self.drop_path = DropPath(drop_path_rate) if drop_path_rate > 0. else nn.Identity()
         self.act = nn.Identity() if linear_out else layers.act(inplace=True)
 
-    def init_weights(self, zero_init_last_bn: bool = False):
-        if zero_init_last_bn:
+    def init_weights(self, zero_init_last: bool = False):
+        if zero_init_last:
             nn.init.zeros_(self.conv2_kxk.bn.weight)
         for attn in (self.attn, self.attn_last):
             if hasattr(attn, 'reset_parameters'):
@@ -814,8 +814,8 @@ def __init__(self, in_chs, out_chs, kernel_size=3, stride=1, dilation=(1, 1), bo
         self.drop_path = DropPath(drop_path_rate) if drop_path_rate > 0. else nn.Identity()
         self.act = nn.Identity() if linear_out else layers.act(inplace=True)
 
-    def init_weights(self, zero_init_last_bn: bool = False):
-        if zero_init_last_bn:
+    def init_weights(self, zero_init_last: bool = False):
+        if zero_init_last:
             nn.init.zeros_(self.conv3_1x1.bn.weight)
         for attn in (self.attn, self.attn_last):
             if hasattr(attn, 'reset_parameters'):
@@ -871,8 +871,8 @@ def __init__(self, in_chs, out_chs, kernel_size=3, stride=1, dilation=(1, 1), bo
         self.drop_path = DropPath(drop_path_rate) if drop_path_rate > 0. else nn.Identity()
         self.act = nn.Identity() if linear_out else layers.act(inplace=True)
 
-    def init_weights(self, zero_init_last_bn: bool = False):
-        if zero_init_last_bn:
+    def init_weights(self, zero_init_last: bool = False):
+        if zero_init_last:
             nn.init.zeros_(self.conv2_kxk.bn.weight)
         for attn in (self.attn, self.attn_last):
             if hasattr(attn, 'reset_parameters'):
@@ -924,8 +924,8 @@ def __init__(self, in_chs, out_chs, kernel_size=3, stride=1, dilation=(1, 1), bo
         self.drop_path = DropPath(drop_path_rate) if drop_path_rate > 0. else nn.Identity()
         self.act = nn.Identity() if linear_out else layers.act(inplace=True)
 
-    def init_weights(self, zero_init_last_bn: bool = False):
-        if zero_init_last_bn:
+    def init_weights(self, zero_init_last: bool = False):
+        if zero_init_last:
             nn.init.zeros_(self.conv2_1x1.bn.weight)
         for attn in (self.attn, self.attn_last):
             if hasattr(attn, 'reset_parameters'):
@@ -967,7 +967,7 @@ def __init__(self, in_chs, out_chs, kernel_size=3, stride=1, dilation=(1, 1), bo
         self.drop_path = DropPath(drop_path_rate) if drop_path_rate > 0. and use_ident else nn.Identity()
         self.act = layers.act(inplace=True)
 
-    def init_weights(self, zero_init_last_bn: bool = False):
+    def init_weights(self, zero_init_last: bool = False):
         # NOTE this init overrides that base model init with specific changes for the block type
         for m in self.modules():
             if isinstance(m, nn.BatchNorm2d):
@@ -1024,8 +1024,8 @@ def __init__(self, in_chs, out_chs, kernel_size=3, stride=1, dilation=(1, 1), bo
         self.drop_path = DropPath(drop_path_rate) if drop_path_rate > 0. else nn.Identity()
         self.act = nn.Identity() if linear_out else layers.act(inplace=True)
 
-    def init_weights(self, zero_init_last_bn: bool = False):
-        if zero_init_last_bn:
+    def init_weights(self, zero_init_last: bool = False):
+        if zero_init_last:
             nn.init.zeros_(self.conv3_1x1.bn.weight)
         if hasattr(self.self_attn, 'reset_parameters'):
             self.self_attn.reset_parameters()
@@ -1278,7 +1278,7 @@ class ByobNet(nn.Module):
     Current assumption is that both stem and blocks are in conv-bn-act order (w/ block ending in act).
     """
     def __init__(self, cfg: ByoModelCfg, num_classes=1000, in_chans=3, global_pool='avg', output_stride=32,
-                 zero_init_last_bn=True, img_size=None, drop_rate=0., drop_path_rate=0.):
+                 zero_init_last=True, img_size=None, drop_rate=0., drop_path_rate=0.):
         super().__init__()
         self.num_classes = num_classes
         self.drop_rate = drop_rate
@@ -1309,12 +1309,8 @@ def __init__(self, cfg: ByoModelCfg, num_classes=1000, in_chans=3, global_pool='
 
         self.head = ClassifierHead(self.num_features, num_classes, pool_type=global_pool, drop_rate=self.drop_rate)
 
-        for n, m in self.named_modules():
-            _init_weights(m, n)
-        for m in self.modules():
-            # call each block's weight init for block-specific overrides to init above
-            if hasattr(m, 'init_weights'):
-                m.init_weights(zero_init_last_bn=zero_init_last_bn)
+        # init weights
+        named_apply(partial(_init_weights, zero_init_last=zero_init_last), self)
 
     def get_classifier(self):
         return self.head.fc
@@ -1334,20 +1330,22 @@ def forward(self, x):
         return x
 
 
-def _init_weights(m, n=''):
-    if isinstance(m, nn.Conv2d):
-        fan_out = m.kernel_size[0] * m.kernel_size[1] * m.out_channels
-        fan_out //= m.groups
-        m.weight.data.normal_(0, math.sqrt(2.0 / fan_out))
-        if m.bias is not None:
-            m.bias.data.zero_()
-    elif isinstance(m, nn.Linear):
-        nn.init.normal_(m.weight, mean=0.0, std=0.01)
-        if m.bias is not None:
-            nn.init.zeros_(m.bias)
-    elif isinstance(m, nn.BatchNorm2d):
-        nn.init.ones_(m.weight)
-        nn.init.zeros_(m.bias)
+def _init_weights(module, name='', zero_init_last=False):
+    if isinstance(module, nn.Conv2d):
+        fan_out = module.kernel_size[0] * module.kernel_size[1] * module.out_channels
+        fan_out //= module.groups
+        module.weight.data.normal_(0, math.sqrt(2.0 / fan_out))
+        if module.bias is not None:
+            module.bias.data.zero_()
+    elif isinstance(module, nn.Linear):
+        nn.init.normal_(module.weight, mean=0.0, std=0.01)
+        if module.bias is not None:
+            nn.init.zeros_(module.bias)
+    elif isinstance(module, nn.BatchNorm2d):
+        nn.init.ones_(module.weight)
+        nn.init.zeros_(module.bias)
+    elif hasattr(module, 'init_weights'):
+        module.init_weights(zero_init_last=zero_init_last)
 
 
 def _create_byobnet(variant, pretrained=False, **kwargs):

timm/models/layers/bottleneck_attn.py

@@ -102,6 +102,8 @@ def __init__(self, dim, dim_out=None, feat_size=None, stride=1, num_heads=4, qkv
 
         self.pool = nn.AvgPool2d(2, 2) if stride == 2 else nn.Identity()
 
+        self.reset_parameters()
+
     def reset_parameters(self):
         trunc_normal_(self.qkv.weight, std=self.qkv.weight.shape[1] ** -0.5)
         trunc_normal_(self.pos_embed.height_rel, std=self.scale)

timm/models/layers/halo_attn.py

@@ -123,6 +123,8 @@ def __init__(
         self.pos_embed = PosEmbedRel(
             block_size=block_size // self.stride, win_size=self.win_size, dim_head=self.dim_head, scale=self.scale)
 
+        self.reset_parameters()
+
     def reset_parameters(self):
         std = self.q.weight.shape[1] ** -0.5  # fan-in
         trunc_normal_(self.q.weight, std=std)

timm/models/layers/lambda_layer.py

@@ -57,6 +57,8 @@ def __init__(
 
         self.pool = nn.AvgPool2d(2, 2) if stride == 2 else nn.Identity()
 
+        self.reset_parameters()
+
     def reset_parameters(self):
         trunc_normal_(self.qkv.weight, std=self.dim ** -0.5)
         trunc_normal_(self.conv_lambda.weight, std=self.dim_k ** -0.5)