Update batchnorm freezing to handle NormAct variants, Add GroupNorm1Act, update BatchNormAct2d tracing change from PyTorch

pull/1633/head
Ross Wightman 1 year ago
parent a2c14c2064
commit e520553e3e

@ -29,7 +29,8 @@ from .mixed_conv2d import MixedConv2d
from .mlp import Mlp, GluMlp, GatedMlp, ConvMlp, GlobalResponseNormMlp from .mlp import Mlp, GluMlp, GatedMlp, ConvMlp, GlobalResponseNormMlp
from .non_local_attn import NonLocalAttn, BatNonLocalAttn from .non_local_attn import NonLocalAttn, BatNonLocalAttn
from .norm import GroupNorm, GroupNorm1, LayerNorm, LayerNorm2d from .norm import GroupNorm, GroupNorm1, LayerNorm, LayerNorm2d
from .norm_act import BatchNormAct2d, GroupNormAct, convert_sync_batchnorm from .norm_act import BatchNormAct2d, GroupNormAct, GroupNorm1Act, LayerNormAct, LayerNormAct2d,\
SyncBatchNormAct, convert_sync_batchnorm, FrozenBatchNormAct2d, freeze_batch_norm_2d, unfreeze_batch_norm_2d
from .padding import get_padding, get_same_padding, pad_same from .padding import get_padding, get_same_padding, pad_same
from .patch_embed import PatchEmbed, resample_patch_embed from .patch_embed import PatchEmbed, resample_patch_embed
from .pool2d_same import AvgPool2dSame, create_pool2d from .pool2d_same import AvgPool2dSame, create_pool2d

@ -17,6 +17,7 @@ from typing import Union, List, Optional, Any
import torch import torch
from torch import nn as nn from torch import nn as nn
from torch.nn import functional as F from torch.nn import functional as F
from torchvision.ops.misc import FrozenBatchNorm2d
from .create_act import get_act_layer from .create_act import get_act_layer
from .fast_norm import is_fast_norm, fast_group_norm, fast_layer_norm from .fast_norm import is_fast_norm, fast_group_norm, fast_layer_norm
@ -77,7 +78,7 @@ class BatchNormAct2d(nn.BatchNorm2d):
if self.training and self.track_running_stats: if self.training and self.track_running_stats:
# TODO: if statement only here to tell the jit to skip emitting this when it is None # TODO: if statement only here to tell the jit to skip emitting this when it is None
if self.num_batches_tracked is not None: # type: ignore[has-type] if self.num_batches_tracked is not None: # type: ignore[has-type]
self.num_batches_tracked = self.num_batches_tracked + 1 # type: ignore[has-type] self.num_batches_tracked.add_(1) # type: ignore[has-type]
if self.momentum is None: # use cumulative moving average if self.momentum is None: # use cumulative moving average
exponential_average_factor = 1.0 / float(self.num_batches_tracked) exponential_average_factor = 1.0 / float(self.num_batches_tracked)
else: # use exponential moving average else: # use exponential moving average
@ -169,6 +170,159 @@ def convert_sync_batchnorm(module, process_group=None):
return module_output return module_output
class FrozenBatchNormAct2d(torch.nn.Module):
"""
BatchNormAct2d where the batch statistics and the affine parameters are fixed
Args:
num_features (int): Number of features ``C`` from an expected input of size ``(N, C, H, W)``
eps (float): a value added to the denominator for numerical stability. Default: 1e-5
"""
def __init__(
self,
num_features: int,
eps: float = 1e-5,
apply_act=True,
act_layer=nn.ReLU,
inplace=True,
drop_layer=None,
):
super().__init__()
self.eps = eps
self.register_buffer("weight", torch.ones(num_features))
self.register_buffer("bias", torch.zeros(num_features))
self.register_buffer("running_mean", torch.zeros(num_features))
self.register_buffer("running_var", torch.ones(num_features))
self.drop = drop_layer() if drop_layer is not None else nn.Identity()
act_layer = get_act_layer(act_layer) # string -> nn.Module
if act_layer is not None and apply_act:
act_args = dict(inplace=True) if inplace else {}
self.act = act_layer(**act_args)
else:
self.act = nn.Identity()
def _load_from_state_dict(
self,
state_dict: dict,
prefix: str,
local_metadata: dict,
strict: bool,
missing_keys: List[str],
unexpected_keys: List[str],
error_msgs: List[str],
):
num_batches_tracked_key = prefix + "num_batches_tracked"
if num_batches_tracked_key in state_dict:
del state_dict[num_batches_tracked_key]
super()._load_from_state_dict(
state_dict, prefix, local_metadata, strict, missing_keys, unexpected_keys, error_msgs
)
def forward(self, x: torch.Tensor) -> torch.Tensor:
# move reshapes to the beginning
# to make it fuser-friendly
w = self.weight.reshape(1, -1, 1, 1)
b = self.bias.reshape(1, -1, 1, 1)
rv = self.running_var.reshape(1, -1, 1, 1)
rm = self.running_mean.reshape(1, -1, 1, 1)
scale = w * (rv + self.eps).rsqrt()
bias = b - rm * scale
x = x * scale + bias
x = self.act(self.drop(x))
return x
def __repr__(self) -> str:
return f"{self.__class__.__name__}({self.weight.shape[0]}, eps={self.eps}, act={self.act})"
def freeze_batch_norm_2d(module):
"""
Converts all `BatchNorm2d` and `SyncBatchNorm` or `BatchNormAct2d` and `SyncBatchNormAct2d` layers
of provided module into `FrozenBatchNorm2d` or `FrozenBatchNormAct2d` respectively.
Args:
module (torch.nn.Module): Any PyTorch module.
Returns:
torch.nn.Module: Resulting module
Inspired by https://github.com/pytorch/pytorch/blob/a5895f85be0f10212791145bfedc0261d364f103/torch/nn/modules/batchnorm.py#L762
"""
res = module
if isinstance(module, (BatchNormAct2d, SyncBatchNormAct)):
res = FrozenBatchNormAct2d(module.num_features)
res.num_features = module.num_features
res.affine = module.affine
if module.affine:
res.weight.data = module.weight.data.clone().detach()
res.bias.data = module.bias.data.clone().detach()
res.running_mean.data = module.running_mean.data
res.running_var.data = module.running_var.data
res.eps = module.eps
res.drop = module.drop
res.act = module.act
elif isinstance(module, (torch.nn.modules.batchnorm.BatchNorm2d, torch.nn.modules.batchnorm.SyncBatchNorm)):
res = FrozenBatchNorm2d(module.num_features)
res.num_features = module.num_features
res.affine = module.affine
if module.affine:
res.weight.data = module.weight.data.clone().detach()
res.bias.data = module.bias.data.clone().detach()
res.running_mean.data = module.running_mean.data
res.running_var.data = module.running_var.data
res.eps = module.eps
else:
for name, child in module.named_children():
new_child = freeze_batch_norm_2d(child)
if new_child is not child:
res.add_module(name, new_child)
return res
def unfreeze_batch_norm_2d(module):
"""
Converts all `FrozenBatchNorm2d` layers of provided module into `BatchNorm2d`. If `module` is itself and instance
of `FrozenBatchNorm2d`, it is converted into `BatchNorm2d` and returned. Otherwise, the module is walked
recursively and submodules are converted in place.
Args:
module (torch.nn.Module): Any PyTorch module.
Returns:
torch.nn.Module: Resulting module
Inspired by https://github.com/pytorch/pytorch/blob/a5895f85be0f10212791145bfedc0261d364f103/torch/nn/modules/batchnorm.py#L762
"""
res = module
if isinstance(module, FrozenBatchNormAct2d):
res = BatchNormAct2d(module.num_features)
if module.affine:
res.weight.data = module.weight.data.clone().detach()
res.bias.data = module.bias.data.clone().detach()
res.running_mean.data = module.running_mean.data
res.running_var.data = module.running_var.data
res.eps = module.eps
res.drop = module.drop
res.act = module.act
elif isinstance(module, FrozenBatchNorm2d):
res = torch.nn.BatchNorm2d(module.num_features)
if module.affine:
res.weight.data = module.weight.data.clone().detach()
res.bias.data = module.bias.data.clone().detach()
res.running_mean.data = module.running_mean.data
res.running_var.data = module.running_var.data
res.eps = module.eps
else:
for name, child in module.named_children():
new_child = unfreeze_batch_norm_2d(child)
if new_child is not child:
res.add_module(name, new_child)
return res
def _num_groups(num_channels, num_groups, group_size): def _num_groups(num_channels, num_groups, group_size):
if group_size: if group_size:
assert num_channels % group_size == 0 assert num_channels % group_size == 0
@ -179,10 +333,54 @@ def _num_groups(num_channels, num_groups, group_size):
class GroupNormAct(nn.GroupNorm): class GroupNormAct(nn.GroupNorm):
# NOTE num_channel and num_groups order flipped for easier layer swaps / binding of fixed args # NOTE num_channel and num_groups order flipped for easier layer swaps / binding of fixed args
def __init__( def __init__(
self, num_channels, num_groups=32, eps=1e-5, affine=True, group_size=None, self,
apply_act=True, act_layer=nn.ReLU, inplace=True, drop_layer=None): num_channels,
num_groups=32,
eps=1e-5,
affine=True,
group_size=None,
apply_act=True,
act_layer=nn.ReLU,
inplace=True,
drop_layer=None,
):
super(GroupNormAct, self).__init__( super(GroupNormAct, self).__init__(
_num_groups(num_channels, num_groups, group_size), num_channels, eps=eps, affine=affine) _num_groups(num_channels, num_groups, group_size),
num_channels,
eps=eps,
affine=affine,
)
self.drop = drop_layer() if drop_layer is not None else nn.Identity()
act_layer = get_act_layer(act_layer) # string -> nn.Module
if act_layer is not None and apply_act:
act_args = dict(inplace=True) if inplace else {}
self.act = act_layer(**act_args)
else:
self.act = nn.Identity()
self._fast_norm = is_fast_norm()
def forward(self, x):
if self._fast_norm:
x = fast_group_norm(x, self.num_groups, self.weight, self.bias, self.eps)
else:
x = F.group_norm(x, self.num_groups, self.weight, self.bias, self.eps)
x = self.drop(x)
x = self.act(x)
return x
class GroupNorm1Act(nn.GroupNorm):
def __init__(
self,
num_channels,
eps=1e-5,
affine=True,
apply_act=True,
act_layer=nn.ReLU,
inplace=True,
drop_layer=None,
):
super(GroupNorm1Act, self).__init__(1, num_channels, eps=eps, affine=affine)
self.drop = drop_layer() if drop_layer is not None else nn.Identity() self.drop = drop_layer() if drop_layer is not None else nn.Identity()
act_layer = get_act_layer(act_layer) # string -> nn.Module act_layer = get_act_layer(act_layer) # string -> nn.Module
if act_layer is not None and apply_act: if act_layer is not None and apply_act:
@ -204,8 +402,15 @@ class GroupNormAct(nn.GroupNorm):
class LayerNormAct(nn.LayerNorm): class LayerNormAct(nn.LayerNorm):
def __init__( def __init__(
self, normalization_shape: Union[int, List[int], torch.Size], eps=1e-5, affine=True, self,
apply_act=True, act_layer=nn.ReLU, inplace=True, drop_layer=None): normalization_shape: Union[int, List[int], torch.Size],
eps=1e-5,
affine=True,
apply_act=True,
act_layer=nn.ReLU,
inplace=True,
drop_layer=None,
):
super(LayerNormAct, self).__init__(normalization_shape, eps=eps, elementwise_affine=affine) super(LayerNormAct, self).__init__(normalization_shape, eps=eps, elementwise_affine=affine)
self.drop = drop_layer() if drop_layer is not None else nn.Identity() self.drop = drop_layer() if drop_layer is not None else nn.Identity()
act_layer = get_act_layer(act_layer) # string -> nn.Module act_layer = get_act_layer(act_layer) # string -> nn.Module
@ -228,8 +433,15 @@ class LayerNormAct(nn.LayerNorm):
class LayerNormAct2d(nn.LayerNorm): class LayerNormAct2d(nn.LayerNorm):
def __init__( def __init__(
self, num_channels, eps=1e-5, affine=True, self,
apply_act=True, act_layer=nn.ReLU, inplace=True, drop_layer=None): num_channels,
eps=1e-5,
affine=True,
apply_act=True,
act_layer=nn.ReLU,
inplace=True,
drop_layer=None,
):
super(LayerNormAct2d, self).__init__(num_channels, eps=eps, elementwise_affine=affine) super(LayerNormAct2d, self).__init__(num_channels, eps=eps, elementwise_affine=affine)
self.drop = drop_layer() if drop_layer is not None else nn.Identity() self.drop = drop_layer() if drop_layer is not None else nn.Identity()
act_layer = get_act_layer(act_layer) # string -> nn.Module act_layer = get_act_layer(act_layer) # string -> nn.Module

@ -7,6 +7,8 @@ import fnmatch
import torch import torch
from torchvision.ops.misc import FrozenBatchNorm2d from torchvision.ops.misc import FrozenBatchNorm2d
from timm.layers import BatchNormAct2d, SyncBatchNormAct, FrozenBatchNormAct2d,\
freeze_batch_norm_2d, unfreeze_batch_norm_2d
from .model_ema import ModelEma from .model_ema import ModelEma
@ -100,70 +102,6 @@ def extract_spp_stats(
return hook.stats return hook.stats
def freeze_batch_norm_2d(module):
"""
Converts all `BatchNorm2d` and `SyncBatchNorm` layers of provided module into `FrozenBatchNorm2d`. If `module` is
itself an instance of either `BatchNorm2d` or `SyncBatchNorm`, it is converted into `FrozenBatchNorm2d` and
returned. Otherwise, the module is walked recursively and submodules are converted in place.
Args:
module (torch.nn.Module): Any PyTorch module.
Returns:
torch.nn.Module: Resulting module
Inspired by https://github.com/pytorch/pytorch/blob/a5895f85be0f10212791145bfedc0261d364f103/torch/nn/modules/batchnorm.py#L762
"""
res = module
if isinstance(module, (torch.nn.modules.batchnorm.BatchNorm2d, torch.nn.modules.batchnorm.SyncBatchNorm)):
res = FrozenBatchNorm2d(module.num_features)
res.num_features = module.num_features
res.affine = module.affine
if module.affine:
res.weight.data = module.weight.data.clone().detach()
res.bias.data = module.bias.data.clone().detach()
res.running_mean.data = module.running_mean.data
res.running_var.data = module.running_var.data
res.eps = module.eps
else:
for name, child in module.named_children():
new_child = freeze_batch_norm_2d(child)
if new_child is not child:
res.add_module(name, new_child)
return res
def unfreeze_batch_norm_2d(module):
"""
Converts all `FrozenBatchNorm2d` layers of provided module into `BatchNorm2d`. If `module` is itself and instance
of `FrozenBatchNorm2d`, it is converted into `BatchNorm2d` and returned. Otherwise, the module is walked
recursively and submodules are converted in place.
Args:
module (torch.nn.Module): Any PyTorch module.
Returns:
torch.nn.Module: Resulting module
Inspired by https://github.com/pytorch/pytorch/blob/a5895f85be0f10212791145bfedc0261d364f103/torch/nn/modules/batchnorm.py#L762
"""
res = module
if isinstance(module, FrozenBatchNorm2d):
res = torch.nn.BatchNorm2d(module.num_features)
if module.affine:
res.weight.data = module.weight.data.clone().detach()
res.bias.data = module.bias.data.clone().detach()
res.running_mean.data = module.running_mean.data
res.running_var.data = module.running_var.data
res.eps = module.eps
else:
for name, child in module.named_children():
new_child = unfreeze_batch_norm_2d(child)
if new_child is not child:
res.add_module(name, new_child)
return res
def _freeze_unfreeze(root_module, submodules=[], include_bn_running_stats=True, mode='freeze'): def _freeze_unfreeze(root_module, submodules=[], include_bn_running_stats=True, mode='freeze'):
""" """
Freeze or unfreeze parameters of the specified modules and those of all their hierarchical descendants. This is Freeze or unfreeze parameters of the specified modules and those of all their hierarchical descendants. This is
@ -179,7 +117,12 @@ def _freeze_unfreeze(root_module, submodules=[], include_bn_running_stats=True,
""" """
assert mode in ["freeze", "unfreeze"], '`mode` must be one of "freeze" or "unfreeze"' assert mode in ["freeze", "unfreeze"], '`mode` must be one of "freeze" or "unfreeze"'
if isinstance(root_module, (torch.nn.modules.batchnorm.BatchNorm2d, torch.nn.modules.batchnorm.SyncBatchNorm)): if isinstance(root_module, (
torch.nn.modules.batchnorm.BatchNorm2d,
torch.nn.modules.batchnorm.SyncBatchNorm,
BatchNormAct2d,
SyncBatchNormAct,
)):
# Raise assertion here because we can't convert it in place # Raise assertion here because we can't convert it in place
raise AssertionError( raise AssertionError(
"You have provided a batch norm layer as the `root module`. Please use " "You have provided a batch norm layer as the `root module`. Please use "
@ -213,13 +156,18 @@ def _freeze_unfreeze(root_module, submodules=[], include_bn_running_stats=True,
# It's possible that `m` is a type of BatchNorm in itself, in which case `unfreeze_batch_norm_2d` won't # It's possible that `m` is a type of BatchNorm in itself, in which case `unfreeze_batch_norm_2d` won't
# convert it in place, but will return the converted result. In this case `res` holds the converted # convert it in place, but will return the converted result. In this case `res` holds the converted
# result and we may try to re-assign the named module # result and we may try to re-assign the named module
if isinstance(m, (torch.nn.modules.batchnorm.BatchNorm2d, torch.nn.modules.batchnorm.SyncBatchNorm)): if isinstance(m, (
torch.nn.modules.batchnorm.BatchNorm2d,
torch.nn.modules.batchnorm.SyncBatchNorm,
BatchNormAct2d,
SyncBatchNormAct,
)):
_add_submodule(root_module, n, res) _add_submodule(root_module, n, res)
# Unfreeze batch norm # Unfreeze batch norm
else: else:
res = unfreeze_batch_norm_2d(m) res = unfreeze_batch_norm_2d(m)
# Ditto. See note above in mode == 'freeze' branch # Ditto. See note above in mode == 'freeze' branch
if isinstance(m, FrozenBatchNorm2d): if isinstance(m, (FrozenBatchNorm2d, FrozenBatchNormAct2d)):
_add_submodule(root_module, n, res) _add_submodule(root_module, n, res)

Loading…
Cancel
Save