I was trying to install PyTorch in Coral Dev Board. I could not find any online reference for the installation.
Can I install PyTorch Mobile (formerly Lite) in Coral Dev Board?
Does the architecture support the installation?
I'm not an expert on pytorch but I don't believe they currently have any releases for arm platforms. Although people has been able to build it from source and get it working on arm platforms like jetson nano. This is probably the best guide that you can follow or based from in order figure out the installation:
https://gist.github.com/dusty-nv/ef2b372301c00c0a9d3203e42fd83426
I modified it slightly to fit the dev board but haven't tried yet:
# clone pyTorch repo
git clone http://github.com/pytorch/pytorch
cd pytorch
git submodule update --init --recursive
# install prereqs
sudo pip3 install -U setuptools
sudo pip3 install -r requirements.txt
# Develop Mode:
python3 setup.py install
FYI, I do want to mention that the edgetpu is specialized for processing tflite tensor only. I just wanted to make sure you know that before going through with the installation and realized that there isn't any performance boost when using the pytorch API.
Related
I have trouble in installing PCL 1.9 on my Ubuntu 18.04. Could anybody please help? Really thanks.
I already tried as many tutorials on the Internet. However they both won't work.
When trying to add ppa source and use apt/apt-get to install libpcl-all, it seems that the source doesn't work for ubuntu18.
When I was trying to build the PCL myself, on Ubuntu 18.04, it has lots of dependencies problems. Many tutorials say that using apt to install the dependencies, however some libraries are not available in apt.
There are some people suggesting to use apt install libpcl-dev. Although there is no errors in installation, when I tried to compile an example code, it still doesn't work.
Using pcl-trunk might be your best choice.
git clone https://github.com/PointCloudLibrary/pcl pcl-trunk
cd pcl-trunk && mkdir build && cd build
cmake ..
make
sudo make install
How can I have python3.6 in tensorflow docker images.
All the images I tried (latest, nighty) are using python3.5 and I don't want to modify all my scripts.
The Tensorflow images are based on Ubuntu 16.04, as you can see from the Dockerfile. This release ships with Python 3.5 as standard.
So you'll have to re-build the image, and the Dockerfile will need editing, even though you need to do the actual build with the parameterized_docker_build.sh script.
This answer on ask Ubuntu covers how to get Python 3.6 on Ubuntu 16.04
The simplest way would probably be just to change the From line in the Dockerfile to FROM ubuntu:16.10, and python to python3.6 in the initial apt-get install line
Of course, this may break some other Ubuntu version-specific thing, so an alternative would be to keep Ubuntu 16.04 and install one of the alternative ppa's also listed in the linked answer:
RUN add-apt-repository ppa:deadsnakes/ppa &&
apt-get update &&
apt-get install -y python3.6
Note that you'll need this after the initial apt-get install, because that installs software-properties-common, which you need to add the ppa.
Note also, as in the comments to the linked answer, that you will need to symlink to Python 3.6.
Finally, note that I haven't tried any of this. The may be gotchas, and you may need to make another change to ensure that the correct version of Python is used by the running container.
You can use stable images which are supplied by third parties, like ufoym/deepo.
One that fits TensorFlow, python3.6 and cuda10 can be found here or you can pull it directly using the command docker pull ufoym/deepo:py36-cu100
I use their images all the time, never had problems
With this anwer, I just wanted to specify how I solved this problem (the previous answer of SiHa helped me a lot but I had to add a few steps so that it worked completly).
Context:
I'm using a package (segmentation model for unet++) that requires tensorflow==1.4.0 and keras==2.2.2.
I tried to use the docker image for tensorflow 1.4.0, however, the default version of python of this image is 3.5 which is not compatible with my package.
I managed to install python3.6 on the docker images thanks to the following files:
My Dockerfile contains the following lines:
Dockerfile:
FROM tensorflow/tensorflow:1.4.0-gpu-py3
RUN mkdir /AI_PLATFORM
WORKDIR /AI_PLATFORM
COPY ./install.sh ./install.sh
COPY ./requirements.txt ./requirements.txt
COPY ./computer_vision ./computer_vision
COPY ./config.ini ./config.ini
RUN bash install.sh
Install.sh:
#!/urs/bin/env bash
pip install --upgrade pip
apt-get update
apt-get install -y python3-pip
add-apt-repository ppa:deadsnakes/ppa &&
apt-get update &&
apt-get install python3.6 --assume-yes
apt-get install libpython3.6
python3.6 -m pip install --upgrade pip
python3.6 -m pip install -r requirements.txt
Three things are important:
use python3.6 -m pip instead of pip, else the packages are installed on python 3.5 default version of Ubuntu 16.04
use docker run python3.6 <command> to run your containers with python==3.6
in the requirements.txt file, I had to specify the following things:
h5py==2.10.0
tensorflow-gpu==1.4.1
keras==2.2.2
keras-applications==1.0.4
keras-preprocessing==1.0.2
I hope that this answer will be useful
Maybe the image I created will help you. It is based on the cuda-10.0-devel image and has tensorflow 2.0a-gpu installed.
You can use it as base image for your own implementation. The image itself doesn't do anything. I put the image on dockerhub https://cloud.docker.com/repository/docker/patientzero/tensorflow2.0a-gpu-py3.6
The github repo is located here: https://github.com/patientzero/tensorflow2.0-python3.6-Docker
Pulling it won't do much, but for completeness:
$ docker pull patientzero/tensorflow2.0-gpu-py3.6
edit: changed to general tensorflow 2.0x image.
Also as mentioned here, the official image for the beta 2.0 release now comes with python 3.6 support
A friend and I are interested in training the tesseract-OCR engine for a CV project. We tried using some wrappers such as PyTesser and pyocr, but the results are currently not as accurate as we need them to be. As such, we want to try training the tesseract to perform better for our purposes (i.e. identifying text on food labels), but are having some trouble installing the training tools.
What we've tried:
Looking on the google code website, the 'Compiling' page on the tesseract's google code wiki says the training tools are only available on version 3.03. However, the google code 'Downloads' page for tesseract-ocr only has the materials for 3.02. The bottom of the 'Compiling' page also has some comments about installing version 3.03 on Windows and OSX, but no comments yet for Linux users.
There also appears to be some sort of 3.03 source package for Ubuntu but we're not sure how to access it on our computers and the 'Compiling' page says we need to run these commands:
make training
sudo make training-install
We've also found a google group thread about tesseract 3.03 but again it seems like these posts do not include advice for Linux users (unless we missed something during the initial read).
Is this actually a really simple command-line install problem? Or, is there a way train tesseract with 3.02 (which we currently have installed)? Have we been looking at the wrong places for information?
Any advice or links to instructions for installing tesseract-ocr 3.03 for Linux distributions would be greatly appreciated! Thanks.
Tesseract can directly be installed in Ubuntu 14.04 using
sudo apt-get install tesseract-ocr
I don't have any idea if you can do it in older version of Ubuntu because the repo might be updated in later version of Ubuntu.
I had an aws ubuntu 14.04 instance.
when I tried installing Tesseract with
sudo apt-get install tesseract-ocr
It retuned package not found
But this worked for me.
sudo apt-get update
sudo apt-get install tesseract-ocr
Ubuntu is a debian based Linux distribution. The tesseract package you find will most likely be a debian package which will contain tesseract and the required default language files to allow you to run/train tesseract. You do NOT want the source package -- unless you just want to compile it yourself -- no need. You will not have to build tesseract, you just need to install the package. First, it appears you are new to Ubuntu, so please ready InstallingSoftware. It can be as easy as opening up an x-term and issuing the command apt-get install tesseract-pkgname (note: that means whatever the package name is).
There is no shortcut, take the time to understand whether you have a .deb package on your box that need to be installed or whether you are installing from a remote repository. The link above explains how to handle both.
Here is a specific Ubuntu thread dealing with installing tesseract Tesseract 3.0 + Ubuntu 10.04 Installation Guide Hope that helps. Tesseract is very good software.
I don't have any instructions for building Tesseract 3.03 for Linux specifically (I'm on Mac), but here's a link to download the source code for the 3.03 release candidate: https://tesseract-ocr.googlecode.com/archive/3.03-rc1.tar.gz
First run below command
sudo apt-get install tesseract-ocr
It will install tesseract version 3.04
Run below to update the tesseract
sudo apt-get --only-upgrade install tesseract-ocr
It will update tesseract to 4.1.3
I'm trying to install SciKit learn on a Red Hat Server. According to documentation on Scikit-Learn's website, I can run the following command on Red Hat to install the the dependencies.
sudo yum -y install gcc gcc-c++ numpy python-devel scipy
However, I don't have root privileges, so I am wondering if I can some how modify the above command to run the command?
No, you cannot modify the package base of your system since you are not root and not in sudoers list.
However, you can try to build that packages from sources, but I'm sure that you will come across a lack of *devel packages but you will not manage to install them due to the same reason.
I'm currently migrating to new computer and I need to reinstall the software I am using which are:
Python 3.3,
Lighttpd (newest version),
Pymongo (newest version),
Ubuntu 12.04 Desktop (The System I'm using)
I started to install Python 3.3 by downloading it from the its official website (in tar.bz2 file) and by following this tutorial. Afterwards I installed Lighttpd and changed the lighttpd.conf for Python by following this tutorial, too.
I tried several paths for my cgi.assign, none of them worked. Especially /opt/python3.3/bin/python3.3 should be working, but it shows 500 - internal Server error all the time with a "hello world" test script.
Now regardless to this problem I have no clue on installing Pymongo. If I try to intall pip OR easy_install python3.3 I have to manually download it and execute the setup.py with my python3.3 executable, right? Because this always fails with an error:
`Error missing zlib on a bundle called distribute-0.7.3 (is this even the right tool I need, because it seems to be a legacy wrapper !?) or unknown url type: https for pymongo2.6.2 itself.`
I'm getting crazy with this setup. Why is this so difficult to handle? Other programs are just a few clicks to install even on a system like Ubuntu, but these particular development tools seem to be really difficult to install.If anybody has an idea on how to install all three together or has information on a better solution please help me out.
The system is used to program Python scripts in Eclipse and trying them out directly on the system (lighttpd). The database used is MongoDB. Python and MongoDB are communicating over the Pymongo driver. I am planning to use the system on a Server distribution on release and it has to be nicely scalable on a high amount of excecutions.
Thanks for your time,
It's easiest to use the Ubuntu repositories:
sudo apt-get update
sudo apt-get install python3 python3-pip lightppd python-pymongo
Or if that only installs the python2.x pymongo, use pip, which you've just installed:
sudo pip-3.3 install pymongo
Or better yet, use a virtualenv with the help of virtualenvwrapper (docs)
sudo pip install virtualenvwrapper
... # follow instructions for installing virtualenvwrapper
mkvirtualenv --python=/usr/bin/python3 -i pymongo mongoppd
workon mongoppd
... which will segregate the environment I've called 'mongoppd' from the rest of your system so you can't cause any trouble. Then you don't need sudo to pip-3.3 install things, just workon mongoppd then pip-3.3 install [...]. Or after the -i flag when you create the virtualenv to get it installed straight away.
In general, on Ubuntu, you should hardly ever have to install something manually. Your first attempt should be using sudo apt-get install (use tab-complete to see what's available or just google "ubuntu 12.04 packages [...]" and you'll find the list of packages). Then for python use pip install or pip-3.3 install as appropriate. You'll only need to run python setup.py install if you need to install a development version of a package or something obscure that's not on pip. I don't think there's a good reason to ever use easy_install these days.