mirrors
/
pytorch-image-models
mirror of https://github.com/huggingface/pytorch-image-models.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

Slight improvement in EfficientNet-B2 native PyTorch weights

pull/16/head
Ross Wightman 2019-06-29 16:17:06 -07:00
parent b8762cc67d
commit c3287aafb3
2 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
README.md
View File

@ -87,7 +87,7 @@ I've leveraged the training scripts in this repository to train a few of the mod
#### @ 260x260 #### @ 260x260
|Model | Prec@1 (Err) | Prec@5 (Err) | Param # | Image Scaling | |Model | Prec@1 (Err) | Prec@5 (Err) | Param # | Image Scaling |
|---|---|---|---|---| |---|---|---|---|---|
| efficientnet_b2 | 79.668 (20.332) | 94.634 (5.366) | 9.11M | bicubic | | efficientnet_b2 | 79.760 (20.240) | 94.714 (5.286) | 9.11M | bicubic |
### Ported Weights ### Ported Weights

2
timm/models/gen_efficientnet.py
View File

@ -84,7 +84,7 @@ default_cfgs = {
url='https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/efficientnet_b1-533bc792.pth', url='https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/efficientnet_b1-533bc792.pth',
input_size=(3, 240, 240), pool_size=(8, 8), interpolation='bicubic', crop_pct=0.882), input_size=(3, 240, 240), pool_size=(8, 8), interpolation='bicubic', crop_pct=0.882),
'efficientnet_b2': _cfg( 'efficientnet_b2': _cfg(
url='https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/efficientnet_b2-d4105846.pth', url='https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/efficientnet_b2-cf78dc4d.pth',
input_size=(3, 260, 260), pool_size=(9, 9), interpolation='bicubic', crop_pct=0.890), input_size=(3, 260, 260), pool_size=(9, 9), interpolation='bicubic', crop_pct=0.890),
'efficientnet_b3': _cfg( 'efficientnet_b3': _cfg(
url='', input_size=(3, 300, 300), pool_size=(10, 10), crop_pct=0.904), url='', input_size=(3, 300, 300), pool_size=(10, 10), crop_pct=0.904),