Changelog¶
V0.2.1¶
New Features
Support
transforms.Grayscale
andtransforms.Solarization
transforms #220Support TorchHub functionality #258
Support
transforms.RandomAffine
method #259Support
transforms.RandomRotation
method #261
Bug Fixes
Improvements
refine TODO items and align transforms.ColorJitter to torchvision(0.13.1):#264
Docs Update
Contributors A total of 3 developers contributed to this release. Thanks @BBuf, @rentainhe, @Flowingsun007
V0.2.0¶
New Features
Support PoolFormer model and pretrained weight #137
Support
IResNet
model for face recognition #160Support Dynamic convolution module #166
Support
transforms.RandomGrayscale
method #171Support
transforms.GaussianBlur
method #188Support
SUN397
,Country211
, dataset #215Support
Flowers102
,FGVCAircraft
,OxfordIIITPet
,DTD
,Food101
,RenderedSST2
,StanfordCars
,PCAM
,EuroSAT
,GTSRB
,CLEVR
,FER2013
dataset #217
Bug Fixes
Fix benchmark normalize mode error #146
Improvements
Docs Update
Contributors A total of x developers contributed to this release.
v0.1.0 (10/02/2022)¶
New Features
Support
trunc_normal_
inflowvision.layers.weight_init
#92Support
PolyLRScheduler
andTanhLRScheduler
inflowvision.scheduler
#85Add
resmlp_12_224_dino
model and pretrained weight #128
Bug Fixes
Fix
F.normalize
usage in SSD #116Fix bug in
EfficientNet
andRes2Net
#122Fix error pretrained weight usage in
vit_small_patch32_384
andres2net50_48w_2s
#128
Improvements
Refator
trunc_normal_
andlinspace
usage in Swin-T, Cross-Former, PVT and CSWin models #100Refator
Vision Transformer
model #115Refine
flowvision.models.ModelCreator
to supportModelCreator.model_list
func #123Refator README #124
Refine
load_state_dict_from_url
inflowvision.models.utils
to support downloading pretrained weights to cache dir~/.oneflow/flowvision_cache
#127Rebuild a cleaner model zoo and test all the model with pretrained weights released in flowvision #128
Docs Update
Update
Vision Transformer
docs #115Add
Getting Started
docs #124Add
resmlp_12_224_dino
docs #128Fix
VGG
docs bug #128Add
ConvNeXt
docs #93
Contributors
A total of 5 developers contributed to this release. Thanks @rentainhe, @simonJJJ, @kaijieshi7, @lixiang007666, @Ldpe2G