Skip to content

Fix bn grad original. #3

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
b-mu wants to merge 2 commits into
base: master
Choose a base branch
from
Open

Fix bn grad original. #3

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
fe6fd26
Add BatchNorm modules to BACKPACK.
b-mu Jun 29, 2021
d2ef3f6
Fix bn grad org.
b-mu Jun 29, 2021
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 2 additions & 2 deletions backpack/__init__.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -83,7 +83,7 @@ def hook_store_io(module, input, output):
input: List of input tensors
output: output tensor
"""
if module.training and (isinstance(module, nn.Conv2d) or isinstance(module, nn.Linear)):
if module.training and (isinstance(module, nn.Conv2d) or isinstance(module, nn.Linear) or isinstance(module, nn.BatchNorm1d) or isinstance(module, nn.BatchNorm2d)):
for i in range(len(input)):
setattr(module, "input{}".format(i), input[i])
module.output = output
Expand Down Expand Up @@ -134,7 +134,7 @@ def hook_run_extensions(module, g_inp, g_out):
for backpack_extension in CTX.get_active_exts():
if CTX.get_debug():
print("[DEBUG] Running extension", backpack_extension, "on", module)
if isinstance(module, nn.Conv2d) or isinstance(module, nn.Linear):
if isinstance(module, nn.Conv2d) or isinstance(module, nn.Linear) or isinstance(module, nn.BatchNorm1d) or isinstance(module, nn.BatchNorm2d):
backpack_extension.apply(module, g_inp, g_out)

if not (
Expand Down
10 changes: 5 additions & 5 deletions backpack/extensions/firstorder/fisher_block_eff/batchnorm1d.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,21 +3,21 @@

from torch import einsum, eye, matmul, ones_like, norm
from torch.linalg import inv
import torch

class FisherBlockEffBatchNorm1d(FisherBlockEffBase):
def __init__(self, damping=1.0):
self.damping = damping
super().__init__(derivatives=BatchNorm1dDerivatives(), params=["bias", "weight"])

def weight(self, ext, module, g_inp, g_out, backproped):

return module.weight.grad
update = torch.empty_like(module.weight.grad).copy_(module.weight.grad)
return update


def bias(self, ext, module, g_inp, g_out, backproped):


return module.bias.grad
update = torch.empty_like(module.bias.grad).copy_(module.bias.grad)
return update



10 changes: 6 additions & 4 deletions backpack/extensions/firstorder/fisher_block_eff/batchnorm2d.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,18 +4,20 @@
from torch import einsum, eye, matmul, ones_like, norm
from torch.linalg import inv

import torch

class FisherBlockEffBatchNorm2d(FisherBlockEffBase):
def __init__(self, damping=1.0):
self.damping = damping
super().__init__(derivatives=BatchNorm2dDerivatives(), params=["bias", "weight"])

def weight(self, ext, module, g_inp, g_out, backproped):

return module.weight.grad
update = torch.empty_like(module.weight.grad).copy_(module.weight.grad)
return update


def bias(self, ext, module, g_inp, g_out, backproped):

return module.bias.grad
update = torch.empty_like(module.bias.grad).copy_(module.bias.grad)
return update