GitHub · Where software is built

This repository was archived by the owner on Jul 2, 2023. It is now read-only.

The learning rate is too small
#54 · youngfly11 opened on Sep 22, 2017
5
About VGG16 pretrained model
#52 · zijundeng opened on Aug 29, 2017
6
Why do we need padding=100 for a filter of size 3?
#40 · deartonym opened on Jun 26, 2017
5

Labels Milestones

How to change batch_size？

#168

· Lukashsx opened

on Apr 29, 2023

Gradient explosion occurs after modifying the weight initialization method

#167

· eadstry opened

on Oct 22, 2022

这是怎么处理的？

#165

· Shajiu opened

on Nov 24, 2021

What h[:, :, 31:31 + x.size()[2], 31:31 + x.size()[3]] mean?

#161

· lyw615 opened

on Jun 10, 2021

Pretrained model 404

#160

· BlackDriver opened

on May 30, 2021

Why the padding of first conv layer in FCN32s is 100?

#157

· msy1412 opened

on Nov 29, 2020

requests.exceptions.ConnectionError

#154

· Wenenen opened

on Jun 30, 2020

I use vgg16 pretrained model to train ,but the result of fcn32s is zero, can you help me?

#150

· hylovezhang opened

on Mar 19, 2020

The begining index to crop output images in fcn32s.py

#147

· gniknoil opened

on Feb 9, 2020

Why the parameter of bias is set False to nn.ConvTranspose2d() in fcn8s.py?

#146

· songyuc opened

on Jan 17, 2020

Load FCN8 weights

#145

· luistelmocosta opened

on Dec 4, 2019

I have the same problem, all output results are zeros. Anyone can help ?

#144

· prfans opened

on Nov 21, 2019