Loading...

NVIDIA Jetson TK1 - Caffe Deep Learning Framework

. 52 Videos
Subscribe
22,548,145
2,729,347 views
95% 39,852 852

About :

Install Caffe Deep Learning Framework on Jetson LT4 21.2 with CUDA 6.5. Like and Subscribe if you want more like this.

You can read more about this installation and get the script here:

http://jetsonhacks.com/2015/01/17/nvidia-jetson-tk1-caffe-deep-learning-framework/

In the video, the result for the 'Average Forward Pass' time is a cummulation of 10 iterations in 235 ms, which is around 24 ms per recognition result.

Website: http://jetsonhacks.com
Github: https://github.com/jetsonhacks
Twitter: http://twitter.com/jetsonhacks

annexl option

naruto vs pain dublado pt br

pes 2016 roda em pc fraco

como hackear qualquer jogo

como usar o lucky patcher no clash of clans

serial reset epson l375

erro acesso negado write to disk utorrent

meu moto x nao conecta no pc

comandos craftlandia 1.5.2

como instalar o opl no memory card

Tags :

NVIDIA Jetson TK1 Caffe Deep Learning CUDA 6.5

Google AdSense
728 x 90

Similar Videos:

50 Comments

    Popular Comments:

Jeremy Long Yi . 2017-02-13
the lesson is very great, thanks very much.
121 2 . Reply
Shubhobrata Bhattacharya . 2016-09-14
This doesn't have the provision for GPU activation. Secondly It didn't worked for me in UBUNTU 16.06 with Geforce GT 710. Please help.
121 2 . Reply
Ahmet Can Ozbek . 2016-04-27
What is the backing track of this music?
121 2 . Reply
Saad Qazi . 2016-04-08
Your videos are great - thank you!
121 2 . Reply
Ghassem Tofighi . 2016-02-24
Suppose we have trained model using Jetson or another GPU machine. How we can use this model for detection inside Jetson TK1/TX1 with a good performance? I'm looking for a real-time detection rate (about 30ms for each image). Is it possible at all?!
I tried this example for testing:
http://docs.opencv.org/trunk/d5/de7/tutorial_dnn_googlenet.html
And it takes about 2.7 seconds for each example for GoogleNet on Jetson TX1.
Also tried it for LeNet Mnist and it fails:
http://stackoverflow.com/questions/35465651/how-to-use-opencv-dnn-module-for-digit-classification-by-using-lenet
121 2 . Reply
Phung huynh xuan . 2016-01-23
When I run make -j 4 runtest. It causes the error: cuda driver version is insufficient for cuda runtime version ubuntu. Please help me
121 2 . Reply
顏任佑 . 2016-01-20
Hi! thank for share this video that it make me successful to install caffe in my TK1
But i still have some question to ask
1.Why tk1 can't use LMDB,i try anything i think but LMDB can't work on my tk1
and i need to change every LMDB to leveldb
2.after 1. i creat a caffemodel successful by example mnist,but how can i use this caffemodel?I want to input a image that it can tell me what is that
3.TK1's cpu is arm that i can't install digits ,have another way to show the model or result?

thank you very much!!! Have nice day
121 2 . Reply
Ultimator Korn . 2016-01-01
i have problem with this command "cd caffe && git checkout dev " cd caffe is working fine but git checkout dev does not working properly can you help me please Jetsonhacks ?
121 2 . Reply
glabka333 . 2015-09-20
Cool music, what is a name of the song and band? Thank you.
121 2 . Reply
Wiem Rachman . 2015-08-19
Hi, Can we installed it on Raspberry?.. Pls help!
121 2 . Reply
Audrey Hendon . 2015-06-18
for some reason when I run the line
build/tools/caffe time --model=models/bvlc_alexnet/deploy.prototxt --gpu=0

it will get to the point where it says its going to run 50 iterations, but then it says "killed"

However, when I run the CPU only test, it completes. Any idea what is going on?
121 2 . Reply
游鎮寧 . 2015-06-01
Hi ,I've already installed CUDA6.5 and LT4
but when I command "sudo ./isntallcaffe.sh"
it show "error opencv2/core/core.hpp: No such file or directory"
"installcaffe.sh:/ build/tools/caffe: not found "
Do I need install opencv?
can you help me?
121 2 . Reply
Abhishek Sharma . 2015-05-13
Hi, after make -j 4 all, it shows an error message: cublas not found.. Pls help!
121 2 . Reply
basuam . 2015-05-05
Jetsonhacks...
I'm trying to follow your video but at some point it crashes saying that:

"src/caffe/layers/data_layer.cpp:1:33: fatal error: opencv2/core/core.hpp: No such file or directory
# include <opencv2/core/core.hpp>"

And this is not the only error. How can I add that file and also correct other errors like:

"In file included from ./include/caffe/common.hpp:18:0,
                       from ./include/caffe/blob.hpp:4,
                       from src/caffe/layers/inner_product_layer.cpp:3:
./include/caffe/util/device_alternate.hpp:34:23: fatal error: cublas_v2.h: No such file or directory
# include <cublas_v2.h>"

Hopefully you can answer as soon as possible.  Thank you
121 2 . Reply
GetLost . 2015-05-03
ive got a bit of an issue with this
tried typong
sudo ./installCaffe.sh
about a hundred times now, it keeps saying installcaffe.sh command not found
what am i doing wrong?
121 2 . Reply
赵志昊 . 2015-02-09
Very thanks.Now I have installed caffe on Ubuntu14.04 successfully though I did it for 10 days.But I don't know how to use it... Is there some videos on how to use caffe? 
121 2 . Reply
赵志昊 . 2015-02-02
Is this instruction available for a new-install ubuntu system?   
121 2 . Reply
JetsonHacks . 2015-01-22
jjkjkjka - Where's the fun if you don't make outrageous claims like in the video? Assuming that you can get ~4x speedup CPU only (my guess is that 3x is generous considering overhead) then the GPU mode is still about 6-8x faster. 
121 2 . Reply
jjkjkjka . 2015-01-22
Single threaded CPU speedtest is not fair.
If you use openmp enabled OpenBLAS(for example), you'll get ~4x speedup in cpu mode.
121 2 . Reply