Gunshi Gupta

Hi @liangzu , I'm also starting to implement the same in tensorflow (I have an implementation for the DOAFN network that does the appearance flow part and i'll build on...

For instance, one doubt that i have is, when normally doing deconvolution(/transposed convolution) as in the decoder part of the network, isn't it so that we usually initialize the convolutional...

Hi, I'm not very familiar with pytorch but one thing i think you should look into is the upsampling. I see that you're first upsampling and then doing a transposed...

Another suggestion: The network i originally implemented was the appearance flow network from https://arxiv.org/abs/1605.03557, and tested it on KITTI sequences. This is the original appearance flow network paper , and...

Oh right, I had missed out the explicit upsampling part. Thanks

I had one doubt, @silverbottlep, it'd be great if you can clarify. In the following lines of code from train_tvsn.py (line 228), why is this being done? I originally thought...

Hi, i installed the gfortran compiler through a binary, for OSX, and install gives that error, > uname: illegal option -- o > usage: uname [-amnprsv] > cd lib/5point; /Applications/Xcode.app/Contents/Developer/usr/bin/make...

it continues to build for longer, then ends with the error: > g++ -c -o Decompose.o -O3 -Wall -Wno-unused-result -std=gnu++0x -I../lib/imagelib -I../lib/sfm-driver -I../lib/matrix -I../lib/5point -I../lib/sba-1.5 -I../lib/ann_1.1_char/include -I../include -I/usr/include/eigen3 -D__NO_UI__ -D__BUNDLER__...

I already have it, i get 'warning:jpeg-8d already installed' when i do 'brew install libjpeg'. Or can i force it to build and link to the jpeg library included in...

I copied the libjpeg.dylib to the bin folder and that error is gone, however I get: g++ -c -o BundleCeres.o -O3 -Wall -Wno-unused-result -std=gnu++0x -I../lib/imagelib -I../lib/sfm-driver -I../lib/matrix -I../lib/5point -I../lib/sba-1.5 -I../lib/ann_1.1_char/include...