TigerVersusT

@quic-hitameht @quic-ssiddego [channelPruning.zip](https://github.com/quic/aimet/files/7708055/channelPruning.zip) Thanks for your reply, the zip file contains original [efficientNet-pytorch](https://github.com/lukemelas/EfficientNet-PyTorch), the modified tracable version and a python file, you can reproduce the error by runing it with...

I suggest you consider puring efficientnets yourself, search "puring filters for efficient convnets" on github, it's not so difficult as you may think ---Original--- From: ***@***.***> Date: Fri, Mar 11,...

@luchangli03 Thanks, it works!

原理是一样的吧，这两个博主复现的都是同一本书的同一个算法，这个仓库实用性强点，codeproject那个更贴近原理一些

Same issue

> > > Can you tell us the extraction code please? Sorry, I will fix this right now

> > > Due to regulations, many Chinese developers have no access to google disk, to solve this, I have downloaded them and stored at [baidu disk](https://pan.baidu.com/s/1Y4k-9Tg8q0YZx-FT8jlTRQ) in where all...

simple, just modify dataloading codes by your own, because there seems no APIs about extending user-defined datasets

> Hi @ashutoshmishra1014, @listener17, > > Indeed the last officially supported PyTorch version in Distiller is 1.3.1. At the moment we don't have plans to update it to support later...

Cool, @cgerum , does it fully support commands in original repository? I tried another [fork](https://github.com/MaximIntegratedAI/distiller) but the 1.7 branch seems doesn't support "--export-to-onnx" and "--thinify"