mirrors/pytorch-image-models
1
0
Fork 0
mirror of https://github.com/huggingface/pytorch-image-models.git synced 2025-06-03 15:01:08 +08:00
Code Issues Projects Releases Wiki Activity

Add first 'Normalizer Free' models. nf_regnet_b1 79.3 @ 288x288 test, and nf_resnet50 80.3 @ 256x256 test (80.68 @ 288x288).

Browse Source
This commit is contained in:
Ross Wightman 2021-02-11 13:15:20 -08:00
parent d8e69206be
commit e4de077021
2 changed files with 10 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
README.md
View File

@ -3,6 +3,9 @@
## What's New
### Feb 10, 2021
* First Normalizer-Free model training experiments done,
* nf_resnet50 - 80.68 top-1 @ 288x288, 80.31 @ 256x256
* nf_regnet_b1 - 79.30 @ 288x288, 78.75 @ 256x256
* More model archs, incl a flexible ByobNet backbone ('Bring-your-own-blocks')
* GPU-Efficient-Networks (https://github.com/idstcv/GPU-Efficient-Networks), impl in `byobnet.py`
* RepVGG (https://github.com/DingXiaoH/RepVGG), impl in `byobnet.py`

10
timm/models/nfnet.py
View File

@ -34,17 +34,21 @@ def _dcfg(url='', **kwargs):
**kwargs
}
# FIXME finish
default_cfgs = {
'nf_regnet_b0': _dcfg(url=''),
'nf_regnet_b1': _dcfg(url='', input_size=(3, 240, 240), pool_size=(8, 8)),
'nf_regnet_b1': _dcfg(
url='https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/nf_regnet_b1_256_ra2-ad85cfef.pth',
pool_size=(8, 8), input_size=(3, 256, 256), test_input_size=(3, 288, 288), crop_pct=0.9),
'nf_regnet_b2': _dcfg(url='', input_size=(3, 256, 256), pool_size=(8, 8)),
'nf_regnet_b3': _dcfg(url='', input_size=(3, 272, 272), pool_size=(9, 9)),
'nf_regnet_b4': _dcfg(url='', input_size=(3, 320, 320), pool_size=(10, 10)),
'nf_regnet_b5': _dcfg(url='', input_size=(3, 384, 384), pool_size=(12, 12)),
'nf_resnet26': _dcfg(url='', first_conv='stem.conv'),
'nf_resnet50': _dcfg(url='', first_conv='stem.conv'),
'nf_resnet50': _dcfg(
url='https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/nf_resnet50_ra2-9f236009.pth',
first_conv='stem.conv', pool_size=(8, 8), input_size=(3, 256, 256), crop_pct=0.94),
'nf_resnet101': _dcfg(url='', first_conv='stem.conv'),
'nf_seresnet26': _dcfg(url='', first_conv='stem.conv'),