Installing Tensorflow Windows "Nightly Build" using PIP - jenkins

I need to install a Tensorflow nightly build since the latest stable release has a critical issue that doesn't allow me to work that seems fixed in nightly.
I want to avoid to try to compile myself Tensorflow from source and I have found a guide about installing Nightly build using PIP
Eg this command
pip install http://ci.tensorflow.org/view/Nightly/job/nightly-matrix-cpu/TF_BUILD_IS_OPT=OPT,TF_BUILD_IS_PIP=PIP,TF_BUILD_PYTHON_VERSION=PYTHON3,label=mac-slave/lastSuccessfulBuild/artifact/pip_test/whl/tensorflow-0.11.0rc1-py3-none-any.whl
was for tensorflow 0.11 nightly build for Mac
Looking at this link and link for windows nightly build:
I have tried to adjust the command to install tensorflow 1.0.1 for windows nightly in this way:
pip install http://ci.tensorflow.org/view/Nightly/job/nightly-win/TF_BUILD_CONTAINER_TYPE=CPU,TF_BUILD_IS_OPT=OPT,TF_BUILD_IS_PIP=PIP,TF_BUILD_PYTHON_VERSION=PYTHON3 /lastSuccessfulBuild/artifact/pip_test/whl/tensorflow-1.0.1-cp35-cp35m-win_amd64.whl
with no success
Could someone help me?

I downloaded the latest Nightly Build from Jenkins (https://ci.tensorflow.org/view/Nightly/job/nightly-win/DEVICE=cpu,OS=windows/121/)
I put the file in C:\tensorflow\tensorflow-1.0.1-cp35-cp35m-win_amd64.whl
And was able to successfully install with
pip3 install C:\tensorflow\tensorflow-1.0.1-cp35-cp35m-win_a
md64.whl

Related

Pydrake has issue with python 3 point 9

Is Python3.9 supported?
I got this error with Python3.9:
File "/home/drake/drake/drake-build/install/lib/python3.6/site-packages/pydrake/common/__init__.py", line 8, in <module> from ._module_py import * ModuleNotFoundError: No module named 'pydrake.common._module_py'
There is no "python3.9" folder in .../install/lib.
I am running Ubuntu 18, and I am building Drake from source with latest github commit in master.
EDIT: Can someone explain how exactly Drake sets up pydrake?
It seems it detects the default Python installation somewhere automatically. I tried with a new installation, the default python was 3.8, and I also install:
apt install -y python3.10
Then I followed Drake python setup instructions.
git clone https://github.com/RobotLocomotion/drake.git
mkdir drake-build
cd drake-build
cmake ../drake
make -j
Pydrake only became available in 3.8. How to make it available for 3.10?
The current version of Pydrake (1.11.0) is officially supported on Ubuntu 20.04 with Python 3.8 and Ubuntu 22.04 with Python 3.10 when building from source. However, we recommend that most users use a binary release, and don't try to rebuild Drake from scratch themselves.
There are precompiled wheels at https://pypi.org/project/drake/ aka pip install drake; helpful installation details are at https://drake.mit.edu/pip.html. The wheels when run on Ubuntu support Python versions 3.8, 3.9, 3.10, or 3.11.
For example:
python3 -m venv env
env/bin/pip install --upgrade pip
env/bin/pip install drake
source env/bin/activate
For additional details, see https://drake.mit.edu/installation.html for full instructions and supported versions.
The last version of Pydrake to support Ubuntu 18.04 was v1.1.0 (released in March of 2022). If you need a newer version of Pydrake, you'll need to use a newer version of Ubuntu.

How to inspect package version installed in dockerfile

In my Dockerfile I run:
RUN pip install requests pandas sqlalchemy psycopg2
But how can I find the versions of these packages ? Is there some command line like :
pip list
You should always do version pinning, so specify the exact version number when you install a package (with any package manager) to have reproducable builds otherwise a package might have breaking changes and break your app on the next build.

Electron builder - Define linux version

Im using electron which has already a production env running. For the last months I've been publishing normally with Electron-builder in my ci.yml file:
release-linux:
stage: release
image: electronuserland/builder:14
before_script:
- export ARCH="x64"
- export BUILD_SCRIPT="build:linux"
- apt-get update && apt-get install -y build-essential cmake
- apt install -y libusb-1.0-0-dev libudev-dev
After running the - apt-get update && apt-get install -y build-essential cmake line there is the libc6 package installed which installs with latest (2.35).
My production environment is running on Ubuntu 20.04 LTS. I assume the docker image in which the release is made, is running on a newer version (asuming latest).
The problem is that I have some hardware devices which works correctly with libc6=2.31-0ubuntu9.2 (which is the version that was automatically installed previously by the docker image). Now, for some reason I don't know, the image is pulling the latest version of libc6 and my hardware devices stoped working.
Is there any way to force electron-builder to use a specific Linux/Ubuntu version?
Or maybe use the latest version of ubuntu but downgrading the libc6 version (tried this one but failed). Any instructions on how to do this are apreciated.
Updating Ubuntu's version at production is not an option right now, so I think the solution should be working out with this package/library version.
The problem solved out on its own randomly, but after some research, it was a electron-builder change which was causing the problem.
In this commit changes log we can see that buildpack-deps was changed from buildpack-deps:22.04-curl (not working) to buildpack-deps:focal-curl (working).
That give us a clue to where to look at. Here the official supported tags for Dockerfiles. Here we can see that focal-curl points to 20.04-curl which I suppose have a different output than 20.04 by its own.
Then, looking at the gitlab jobs running with this ci configs, the libc6 dependency worked with different versions in each case.
Our hardware works correctly with libc6=2.31 which is used when focal-curl is set by electron-builder dockerfile, but not with libc6=2.35 which is used by 20.04.
This is not a solution, but it's the cause of the problem. Now working on how to setup the project to use only the correct configuration of electron-builder. Currently accepted to work with same config but being careful when releasing by watching which libc6 version is working with.

Cannot import MSTL from statsmodels

I am currently trying to import MSTL from statsmodels.tsa.seasonal the module of MSTL (https://www.statsmodels.org/devel/generated/statsmodels.tsa.seasonal.MSTL.html) but it returns an ImportError. I have installed statsmodels from conda on MAC M1 2020
I just had the same issue and did some research.
It seems that MSTL is only available on the most recent version of statsmodels: version 0.14.0
If you install statsmodels using conda install -c conda-forge statsmodels,
you will get the statsmodels 0.13.2 version.
(Using a script editor, try searching for 'MSTL' through C:\Users{username}\Anaconda3\Lib\site-packages\statsmodels, or wherever statsmodels is installed on your machine, you will probably not find it)
You'll need to install the most recent version from the latest source on statsmodels's github repository:
www.statsmodels.org/dev/install.html
From the anaconda prompt:
git clone https://github.com/statsmodels/statsmodels.git
pip install git+https://github.com/statsmodels/statsmodels
You will need a C compiler and git installed
For git you can use: conda install -c anaconda git
Be careful as the installation of the newest version may interfere with your other installed python packages.
I would recommend that you use a conda virtual environment for this.

python3-wheel installed, still using legacy 'setup.py' when installing dlib

I've forked dlandon's Zoneminder project in order to build a Docker image using Debian Bullseye instead of Ubuntu, for i386 (I'm installing it into a 32-bit mini PC).
Part of the requiriments is installing dlib (latest is 19.22.0). I've installed the required dependencies (python3-pip, python3-setuptools and python3-wheel, as well as GCC and CMake).
When installing dlib, the terminal outputs:
Collecting dlib
Downloading dlib-19.22.0.tar.gz (7.4 MB)
Using legacy 'setup.py install' for dlib, since package 'wheel' is not installed.
Installing collected packages: dlib
Running setup.py install for dlib: started
...
The relevant part is (even with python3-wheel installed)
Using legacy 'setup.py install' for dlib, since package 'wheel' is not installed.
dlib builts successfully but takes a looong time to build, however. This happens with any packages supossed to use wheel.
I haven't found any (answered) questions about this specific issue, so, if there is something I missed, I would be told about that.
Everything is in the Dockerfile.
try pip install wheel in your env

Resources