Skip to content

Adding device/dtype factory_kwargs to modules and models #2591

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

Merged
merged 21 commits into from
Oct 3, 2025
Merged

Adding device/dtype factory_kwargs to modules and models #2591

Changes from 1 commit
Commits
Show all changes
21 commits
Select commit Hold shift + click to select a range
7560ca5
Adding dd factory_kwargs to modules in timm/layers, initial model WIP…
rwightman Sep 26, 2025
90a35c8
Add dd factory kwargs to eva, resnet
rwightman Sep 27, 2025
325a6cc
Add dd to other ResNet based models, Res2Net, ResNeSt, SKNet
rwightman Sep 27, 2025
b94c221
Add dd factory kwargs to maxxvit and regnet
rwightman Sep 27, 2025
ee751ef
Add dd factory kwargs to nfnet and resnetv2
rwightman Sep 28, 2025
10e7020
dd factory kwargs for fastvit, convnext, mambaout
rwightman Sep 28, 2025
60db539
Add dd factory kwargs to all EfficientNetBuilder models, MobileNet V1…
rwightman Sep 28, 2025
4d19b34
Fix typo for s2d norm
rwightman Sep 28, 2025
f15f7c9
Add dd factory kwargs to byobnet, cspnet, davit, edgenext
rwightman Sep 29, 2025
4c35b78
Add device/dtype factory kwargs to beit, efficientformer*, efficientv…
rwightman Sep 29, 2025
3a85ed4
avg pool should not have been passed dd
rwightman Sep 29, 2025
8cbbf39
Fix DarkStage device kwargs
rwightman Sep 29, 2025
1e172a0
dd kwargs for naflexvit, needs revisit for nn.Parameters
rwightman Sep 29, 2025
a7dc50f
A whack of classic convnets converted with dd factory kwargs. densene…
rwightman Sep 29, 2025
068e6d4
Remove **dd from two inception reset_classifier calls
rwightman Sep 29, 2025
6a3342c
dd factory kwargs added to a bunch of vit/vit-hybrids. cait, coat, co…
rwightman Sep 30, 2025
c7955eb
Add dd factory kwargs to all swin transformers and volo
rwightman Sep 30, 2025
53caeb0
Add some more dd kwarg updates, crossvit, ghostnet, rdnet, repghost, …
rwightman Oct 1, 2025
21b1ae7
More dd factory kwargs updates. hiera, hieradet_sam2, metaformer, mlp…
rwightman Oct 1, 2025
5cadf13
More dd arg conversions. fasternet, gcvit, hgnet, nextvit, starnet, v…
rwightman Oct 1, 2025
d3fdea8
Typing, super(), buffer dtype fixes for timm/layers and timm/models
rwightman Oct 2, 2025
File filter

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Prev Previous commit
Next Next commit
Fix typo for s2d norm
  • Loading branch information
@rwightman
rwightman committed Sep 28, 2025
commit 4d19b341fcfb84f0813e200310f6856dc570777f
4 changes: 2 additions & 2 deletions timm/models/_efficientnet_blocks.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -177,7 +177,7 @@ def __init__(
if s2d == 1:
sd_chs = int(in_chs * 4)
self.conv_s2d = create_conv2d(in_chs, sd_chs, kernel_size=2, stride=2, padding='same', **dd)
self.bn_s2d = norm_act_layer(sd_ch, **dds)
self.bn_s2d = norm_act_layer(sd_chs, **dd)
dw_kernel_size = (dw_kernel_size + 1) // 2
dw_pad_type = 'same' if dw_kernel_size == 2 else pad_type
in_chs = sd_chs
Expand Down Expand Up @@ -275,7 +275,7 @@ def __init__(
if s2d == 1:
sd_chs = int(in_chs * 4)
self.conv_s2d = create_conv2d(in_chs, sd_chs, kernel_size=2, stride=2, padding='same', **dd)
self.bn_s2d = norm_act_layer(sd_chs, sd_chs, **dd)
self.bn_s2d = norm_act_layer(sd_chs, **dd)
dw_kernel_size = (dw_kernel_size + 1) // 2
dw_pad_type = 'same' if dw_kernel_size == 2 else pad_type
in_chs = sd_chs
Expand Down