Closed marek094 closed 6 years ago
Problem does not occur after selecting BitFusion AWS AMI based on Ubuntu 14 from 2016, with properly installed Torch7 and cudnn.
That is good to know. I'm sorry you had to go through this struggle. I will soon release tf based code.
Would it be possible to say more precisely when you are going to be ready with that (maybe pre-)release? [I have sent you an email did you received it? :)].
Hello, I've run tests in gvnn with this result:
`~/gvnn$ th
__ | Torch7 / / ____/ / | Scientific computing for Lua. / / / \/ / / \ | Type ? for help // _// _//// | https://github.com/torch | http://torch.ch
th> require 'gvnn' { VolumetricMaxUnpooling : {...} ConcatTable : {...}
...
TemporalRowConvolution : {...} } [0.0882s] th> dofile('test.lua')
Testing nn with type = cuda
Running 1 test 1/1 Disparity1D_single .................................................. [WAIT]Illegal instruction (core dumped)`
Runned on:
~/gvnn$ uname -a Linux ip-172-31-46-126 4.4.0-1061-aws #70-Ubuntu SMP Fri May 25 21:47:34 UTC 2018 x86_64 x86_64 x86_64 GNU/Linux
With LuaJIT:~/gvnn$ luajit -v LuaJIT 2.1.0-beta1 -- Copyright (C) 2005-2015 Mike Pall. http://luajit.org/