adobe/antialiased-cnns issues and pull requests

#56 - Configure Renovate

Pull Request - State: open - Opened by renovate[bot] 10 months ago

#55 - HTTP Error 403: Forbidden when loading weights

Issue - State: open - Opened by janehwu about 1 year ago - 9 comments

#54 - If stride=1, is there a difference between BlurPool and maxpool

Issue - State: open - Opened by yingmengzhiyin about 1 year ago

#52 - Why do deeper CNNs have better shift consistency?

Issue - State: open - Opened by vadik6666 almost 2 years ago - 2 comments

#51 - Is there any particular reason for puting Blurpool before the skip-connection layers?

Issue - State: open - Opened by freshn almost 2 years ago - 3 comments

#50 - Feature Req: Separable Convolution

Issue - State: open - Opened by torridgristle over 2 years ago

#49 - Feature Req: Making the channel argument optional

Issue - State: open - Opened by torridgristle over 2 years ago

#48 - Could you please provide a 3D implementation in pytorch?

Issue - State: open - Opened by hashemifar almost 3 years ago

#47 - Padding size issue for small images

Issue - State: closed - Opened by DuyguSerbes about 3 years ago - 2 comments

#46 - Any plans to explore using sinc filter for downsampling?

Issue - State: open - Opened by PeterL1n over 3 years ago - 1 comment

#45 - HTTP Error 403: Forbidden when loading weights

Issue - State: closed - Opened by zakajd almost 4 years ago - 2 comments

#44 - Max-blur-pool used in text recognition model (CRNN)

Issue - State: closed - Opened by githubharald almost 4 years ago - 1 comment

#43 - ResNet parameter "pool_only=True"

Issue - State: open - Opened by Yee-Master almost 4 years ago - 1 comment

#42 - Increased memory usage vs. torchvision equivalent

Issue - State: closed - Opened by nlml about 4 years ago - 3 comments

#41 - About Internal feature distance for shift Equivariance visualization

Issue - State: closed - Opened by re20100801 over 4 years ago - 2 comments

#40 - Fix bug in resnet

Pull Request - State: closed - Opened by danielhavir over 4 years ago - 1 comment

#39 - Add 3D blurpooling and code cleanup for 1D and 2D blurpooling

Pull Request - State: open - Opened by nihalsid over 4 years ago - 1 comment

#38 - Add 3D blurpooling and code cleanup for 1D and 2D blurpooling

Pull Request - State: closed - Opened by nihalsid over 4 years ago

#37 - Larger strides/downsampling factors

Issue - State: open - Opened by chanshing over 4 years ago - 2 comments

#36 - 3D implementation ?

Issue - State: open - Opened by romainVala over 4 years ago - 4 comments

#35 - Depthwise convolutions

Issue - State: closed - Opened by aditya1709 over 4 years ago - 1 comment

#34 - can Downsample use tensorflow?

Issue - State: open - Opened by lsjshilly over 4 years ago - 1 comment

#33 - RuntimeError when training resnext50_32x4d

Issue - State: closed - Opened by NightQing over 4 years ago - 2 comments

#32 - Is the blur_kernel trainable or parametric?

Issue - State: closed - Opened by Wuhaotiandl almost 5 years ago - 1 comment

#31 - Can't find Low Pass Filter Usage in Code

Issue - State: open - Opened by humza909 almost 5 years ago - 1 comment

#30 - Bilinear Upsampler Implementation

Issue - State: open - Opened by sr-frost almost 5 years ago - 1 comment

#29 - Why Conv->Relu->BlurPool ?

Issue - State: closed - Opened by mrgloom almost 5 years ago - 1 comment

#28 - What about upsampling?

Issue - State: closed - Opened by chanshing about 5 years ago - 12 comments

#27 - Smooth convolutional filters instead of feature maps?

Issue - State: closed - Opened by mrgloom about 5 years ago - 1 comment

#26 - Is AvrPool already an antialiased downsampling method?

Issue - State: closed - Opened by mrgloom about 5 years ago - 1 comment

#25 - is it possible not to use the pad layer?

Issue - State: closed - Opened by zimenglan-sysu-512 about 5 years ago - 1 comment

#24 - can i cite the figure of the paper directly?

Issue - State: closed - Opened by yelusaleng about 5 years ago - 2 comments

#23 - Add W&B instrumentation

Pull Request - State: closed - Opened by adrnswanberg over 5 years ago

#22 - No strided pooling

Issue - State: closed - Opened by a462428 over 5 years ago - 1 comment

#21 - Asking for Anti-antialiased U-Net codes

Issue - State: closed - Opened by zhangfb over 5 years ago - 2 comments

#20 - Visualizing shift-equivariance

Issue - State: closed - Opened by elphinkuo over 5 years ago - 2 comments

#19 - Some Problems on Reproducing the Results on cub200

Issue - State: open - Opened by yuesongtian over 5 years ago - 2 comments

#18 - Implementation of original ResNet

Issue - State: closed - Opened by jimmy15923 over 5 years ago - 1 comment

#17 - Confuse abount "pool_only" in resnet.py and alexnet.py

Issue - State: closed - Opened by jimmy15923 over 5 years ago

#16 - Unable to establish SSL connection,when use command: bash weights/download_antialiased_models.sh

Issue - State: closed - Opened by LuZeking over 5 years ago - 3 comments

#15 - What about tf.image.resize?

Issue - State: closed - Opened by makercob over 5 years ago - 3 comments

#14 - Add 'IPython' to requirements.txt

Pull Request - State: closed - Opened by adrnswanberg over 5 years ago - 1 comment

#13 - IPython missing from requirements.txt

Issue - State: closed - Opened by adrnswanberg over 5 years ago

#12 - Adaptive version?

Issue - State: closed - Opened by Vlad0922 over 5 years ago - 2 comments

#11 - is it good for autoencoder?

Issue - State: closed - Opened by iperov over 5 years ago - 1 comment

#10 - my keras impl

Issue - State: closed - Opened by iperov over 5 years ago - 18 comments

#9 - FLOPS increase?

Issue - State: closed - Opened by bonlime over 5 years ago - 2 comments

#8 - Are conv2d_transpose layers shift variant, too?

Issue - State: closed - Opened by JuheonYi over 5 years ago - 8 comments

#7 - About the implementation

Issue - State: closed - Opened by boluoweifenda over 5 years ago - 9 comments

#6 - How is this approach related to Fourier Domain Convnets?

Issue - State: closed - Opened by CeaVi over 5 years ago - 1 comment

#5 - About kernel_size of initial max pooing for Resnet

Issue - State: closed - Opened by sweaterr over 5 years ago - 5 comments

#4 - weights connot download can you offer other links,thanks

Issue - State: closed - Opened by dongzhi0312 over 5 years ago - 5 comments

#3 - Utility for own model require a retrained Imagenet Pretrained ?

Issue - State: closed - Opened by JarveeLee over 5 years ago - 5 comments

#2 - Downsampling in 1D

Pull Request - State: closed - Opened by vthorey over 5 years ago - 3 comments

#1 - Use buffers instead of parameters in the Downsample layer

Issue - State: closed - Opened by Skydes over 5 years ago - 2 comments

GitHub / adobe/antialiased-cnns issues and pull requests