-
Hi,
The facenet demo is outdated and not supported anymore by its author, but I will update it when I get to it.
However I will add a more generic demo for object recognition in the next few days. You will be able to use that demo and apply to face tasks if you like.
Please monitor the dev news of the project (https://sourceforge.net/news/?group_id=235450), I will post there when the demo...
2010-01-06 18:34:57 UTC by sermanet
-
hello, I'm also very interested in compiling that facenet eblearn demo, I've found that there is no open discussion on eblearn site, and found this similar question here.
If it not possible to compile eblearn facenet anyway, could you please share a lush or lush2 object detector?
Looking forward to get a reply soon,
Miashiro.
2010-01-04 16:39:20 UTC by sourceforging
-
Hello
I have a question regarding the implementation of facenet (or any other object detector for that matter) using convolutional NNs.
The article (http://eblearn.sourceforge.net/demos/facenet/facenet.shtml) describes the process of creating a detector and references parts of eblearn demo code. However, I had no luck building the demo. The code is already written, the process is...
2009-12-28 22:25:06 UTC by jacobstevens790
-
Thank you for your quick response. I'm trying to make a modified version of you lenet5 code standalone so another program can call it and classify images. I'm trying to use the make-standalone function however, I'm not sure if its even possible to compile the lenet5 code. I'm receiving issues with classifier meter not being a compiled class and the load-matrix button. Maybe I'm going about...
2009-12-26 21:44:35 UTC by pequnio
-
dhc-make is designed to compile Lush functions that asre destined to be used from within Lush.
To produce "standalone" C code from Lush, you must use the "make-standalone" feature. Basically, you define a Lush function (a compilable one), say "myfunc". then you define a "main" function like this:
(de mymain (ac av) ((-int-) ac) ((-gptr- "char **") av) (myfunc))
Say myfunc and mymain...
2009-12-24 01:59:50 UTC by profshadoko
-
You have two ways:
1. do (==> theparam save "somefile"), which you can later reload with (==> theparam load "somefile").
2.do (writing "somefile" (bwrite thetrainer)), which saves the entire thetrainer object.
-- Yann.
2009-12-24 00:39:11 UTC by profshadoko
-
Hi I plan on using a piece of code written in lush and calling it from a c++ program I'm running. Is there a way to do this from c++? Or is it possible for me to compile the code into C. I've looked into using the dhc-make command, but I cant even get it to work for a simple function:
(de foo () (print "hello World"))
(dhc-make () foo)
When I try to compile the .c file that comes out of...
2009-12-23 21:35:22 UTC by pequnio
-
Hello, I've trained a Lenet5 model using the demo code and was wondering if it was possible to save this supervised-gradient "thetrainer" to a file so that it can be read in the future and used to classify test data? Thank you.
2009-12-23 20:35:28 UTC by pequnio
-
rjuengling changed the public information on the Lush project.
2009-12-22 23:05:07 UTC by rjuengling
-
rjuengling changed the public information on the Lush project.
2009-12-22 18:32:19 UTC by rjuengling
