Skip to content
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.

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

ResNet implementation: set bias=False for downsample-B #5477

Merged
merged 5 commits into from
Nov 15, 2022
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
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
13 changes: 11 additions & 2 deletions monai/networks/nets/resnet.py
Original file line number Diff line number Diff line change
Expand Up @@ -175,6 +175,7 @@ class ResNet(nn.Module):
widen_factor: widen output for each layer.
num_classes: number of output (classifications).
feed_forward: whether to add the FC layer for the output, default to `True`.
bias_downsample: whether to use bias term in the downsampling block when `shortcut_type` is 'B', default to `True`.

"""

Expand All @@ -192,6 +193,7 @@ def __init__(
widen_factor: float = 1.0,
num_classes: int = 400,
feed_forward: bool = True,
bias_downsample: bool = True, # for backwards compatibility (also see PR #5477)
) -> None:

super().__init__()
Expand All @@ -216,6 +218,7 @@ def __init__(

self.in_planes = block_inplanes[0]
self.no_max_pool = no_max_pool
self.bias_downsample = bias_downsample

conv1_kernel_size = ensure_tuple_rep(conv1_t_size, spatial_dims)
conv1_stride = ensure_tuple_rep(conv1_t_stride, spatial_dims)
Expand Down Expand Up @@ -277,7 +280,13 @@ def _make_layer(
)
else:
downsample = nn.Sequential(
conv_type(self.in_planes, planes * block.expansion, kernel_size=1, stride=stride),
conv_type(
self.in_planes,
planes * block.expansion,
kernel_size=1,
stride=stride,
bias=self.bias_downsample,
),
norm_type(planes * block.expansion),
)

Expand Down Expand Up @@ -323,7 +332,7 @@ def _resnet(
progress: bool,
**kwargs: Any,
) -> ResNet:
model: ResNet = ResNet(block, layers, block_inplanes, **kwargs)
model: ResNet = ResNet(block, layers, block_inplanes, bias_downsample=not pretrained, **kwargs)
if pretrained:
# Author of paper zipped the state_dict on googledrive,
# so would need to download, unzip and read (2.8gb file for a ~150mb state dict).
Expand Down
18 changes: 17 additions & 1 deletion tests/test_resnet.py
Original file line number Diff line number Diff line change
Expand Up @@ -140,11 +140,27 @@
(1, 3),
]

TEST_CASE_7 = [ # 1D, batch 1, 2 input channels, bias_downsample
{
"block": "bottleneck",
"layers": [3, 4, 6, 3],
"block_inplanes": [64, 128, 256, 512],
"spatial_dims": 1,
"n_input_channels": 2,
"num_classes": 3,
"conv1_t_size": [3],
"conv1_t_stride": 1,
"bias_downsample": False, # set to False if pretrained=True (PR #5477)
},
(1, 2, 32),
(1, 3),
]

TEST_CASES = []
for case in [TEST_CASE_1, TEST_CASE_2, TEST_CASE_3, TEST_CASE_2_A, TEST_CASE_3_A]:
for model in [resnet10, resnet18, resnet34, resnet50, resnet101, resnet152, resnet200]:
TEST_CASES.append([model, *case])
for case in [TEST_CASE_5, TEST_CASE_5_A, TEST_CASE_6]:
for case in [TEST_CASE_5, TEST_CASE_5_A, TEST_CASE_6, TEST_CASE_7]:
TEST_CASES.append([ResNet, *case])

TEST_SCRIPT_CASES = [
Expand Down