I have a running installation of tensorflow-macos and the metal plugin for mac os M1.
Now I want to convert my trained model in a tensorflow js one.
However, when I run
pip install tensorflowjs I get an error due to the fact that tensorflow js actually depends on tensorflow not on tensorflow-macos.
ERROR: Cannot install tensorflowjs==0.1.0, tensorflowjs==0.1.1, tensorflowjs==0.1.2, tensorflowjs==0.2.0, tensorflowjs==0.2.1, tensorflowjs==0.3.0, tensorflowjs==0.3.1, tensorflowjs==0.4.0, tensorflowjs==0.4.1, tensorflowjs==0.4.2, tensorflowjs==0.5.0, tensorflowjs==0.5.2, tensorflowjs==0.5.4, tensorflowjs==0.5.6, tensorflowjs==0.5.7, tensorflowjs==0.6.0, tensorflowjs==0.6.1, tensorflowjs==0.6.2, tensorflowjs==0.6.4, tensorflowjs==0.6.5, tensorflowjs==0.6.7, tensorflowjs==0.8.0, tensorflowjs==0.8.5, tensorflowjs==0.8.6, tensorflowjs==1.0.1, tensorflowjs==1.1.2, tensorflowjs==1.2.1, tensorflowjs==1.2.10, tensorflowjs==1.2.10.1, tensorflowjs==1.2.2, tensorflowjs==1.2.2.1, tensorflowjs==1.2.3, tensorflowjs==1.2.6, tensorflowjs==1.2.9, tensorflowjs==1.3.1, tensorflowjs==1.3.1.1, tensorflowjs==1.3.2, tensorflowjs==1.4.0, tensorflowjs==1.5.2, tensorflowjs==1.6.0, tensorflowjs==1.7.2, tensorflowjs==1.7.3, tensorflowjs==1.7.4, tensorflowjs==1.7.4.post1, tensorflowjs==2.0.0, tensorflowjs==2.0.1, tensorflowjs==2.0.1.post1, tensorflowjs==2.1.0, tensorflowjs==2.3.0, tensorflowjs==2.4.0, tensorflowjs==2.5.0, tensorflowjs==2.6.0, tensorflowjs==2.7.0, tensorflowjs==2.8.0, tensorflowjs==2.8.1, tensorflowjs==2.8.2, tensorflowjs==2.8.3, tensorflowjs==2.8.4, tensorflowjs==2.8.5, tensorflowjs==3.0.0, tensorflowjs==3.1.0, tensorflowjs==3.11.0, tensorflowjs==3.12.0, tensorflowjs==3.13.0, tensorflowjs==3.14.0, tensorflowjs==3.15.0, tensorflowjs==3.17.0, tensorflowjs==3.18.0, tensorflowjs==3.19.0, tensorflowjs==3.2.0, tensorflowjs==3.20.0, tensorflowjs==3.3.0, tensorflowjs==3.4.0, tensorflowjs==3.5.0, tensorflowjs==3.6.0, tensorflowjs==3.7.0, tensorflowjs==3.8.0 and tensorflowjs==3.9.0 because these package versions have conflicting dependencies.
The conflict is caused by:
tensorflowjs 3.20.0 depends on tensorflow<3 and >=2.1.0
...
tensorflowjs 0.1.1 depends on tensorflow>=1.6.0
tensorflowjs 0.1.0 depends on tensorflow>=1.6.0
To fix this you could try to:
1. loosen the range of package versions you've specified
2. remove package versions to allow pip attempt to solve the dependency conflict
Any idea on how to solve it?
just encountered the same problem.
Here is what worked for me.
Install all dependencies of the tensorflowjs package except tensorflow (since you have tensorflow-macos already installed) and then install tensorflowjs without it dependencies.
Install tensorflowjs dependencies :
pip install tensorflow_hub jax scipy jaxlib etils
Install tensorflowjs without its dependencies :
pip install --no-deps tensorflowjs
Related
I'm build up my package with setup.py, in which it has a install_requires=get_requirements('requirements.txt'), and in the requirements.txt, the first required dependency is bottle==0.12.18.
Then I ran
python setup.py sdist
twine upload --repository testpypi dist/*
After uploading succeeded, I installed with pip install -i https://test.pypi.org/simple/ my_package==version_num which gives me the error
Collecting bottle==0.12.18 (from mypakage==version_num)
Could not find a version that satisfies the requirement bottle==0.12.18 (from mypakage==version_num) (from versions: )
No matching distribution found for bottle==0.12.18 (from mypakage==version_num)
Looks like it's looking for the dependency within my package, which will definitely fail. This error suddenly started happening and I've never seen this before. Do you have any idea why it's happening and how can I make it look for dependency in a way like pip install -r requirements?
I have a python package, hosted on azure (vsts), not on pypi, whose dependencies are python packages that live in pypi.
My package has the same name as a package that lives on pypi, I discovered.
Is there a way of installing my package with pip, specifying that my package must be searched on vsts before, while the dependencies can be grabbed from pypi?
If I use the --index-url option:
pip install <my-package> --index-url https://<my-package>:<PAT>#<url>/<proj>/_packaging/<my-package>/pypi/simple/
pip is able to locate my package, tries to install it, but it fails to install any dependency (because it searches for all of them in the same url, which is wrong because I am not hosting, say, my own version of numpy or other packages on vsts).
(This is the problem: pip install producing "Could not find a version that satisfies the requirement" )
If instead I use the --extra-index-url option:
pip install <my-package> --extra-index-url https://<my-package>:<PAT>#<url>/<proj>/_packaging/<my-package>/pypi/simple/
all the dependencies are found, but the problem is that this does not install my package, but the package with the same name that lives in pypi!
Even if --extra-index-url is there, it seems that pypi is given priority, and therefore my package that would be found at the url I specified is shadowed and doesn't get correctly found and installed.
Is there a way to, say, tell pip that it should give priority to my --extra-index-url? Or to give pip an --index-url which should only be valid for one package but not for its dependencies?
You need index URL pointing to VSTS and extra URL to PyPI:
pip install --index-url=https://<my-package>:<PAT>#<url>/<proj>/_packaging/<my-package>/pypi/simple/ --extra-index-url=https://pypi.org/simple/ <my-package>
I'm trying to install the package "Tax4Fun" but keep failing.
I've tried 2 different ways:
install.packages("devtools")
devtools::install_url("http://tax4fun.gobics.de/Tax4Fun/Tax4Fun_0.3.1.tar.gz")
library(Tax4Fun)
The error that I get is:
ERROR: dependency 'biom' is not available for package 'Tax4Fun'
I've also tried installing biom directly
BiocManager::install("biom")
which does not work either
Bioconductor version 3.10 (BiocManager 1.30.10), R 3.6.1 (2019-07-05)
Installing package(s) 'biom'
Installation path not writeable, unable to update packages: boot, foreign, KernSmooth,
mgcv, nlme, survival
Warning message:
package ‘biom’ is not available (for R version 3.6.1)
The other way I've tried to install Tax4Fun directly is
BiocManager::install("Tax4Fun")
I get the following error code:
Bioconductor version 3.10 (BiocManager 1.30.10), R 3.6.1 (2019-07-05)
Installing package(s) 'Tax4Fun'
Installation path not writeable, unable to update packages: boot, foreign, KernSmooth,
mgcv, nlme, survival
Warning message:
package ‘Tax4Fun’ is not available (for R version 3.6.1)
Please help :)
You need to install it by downloading the packages from source (http://tax4fun.gobics.de). Then it depends whether you are running on Linux/Mac or Windows.
From the command line, you navigate to the folder containing the .tar.gz downloaded package. Then you should install it using:
R CMD INSTALL Tax4Fun_0.3.1.tar.gz
But dependancies are not installed by default. So you need to install dependancies manually, Qiimer and Biom, which are both deprecated on Cran. You install them using the same command, after you have downloaded the packages from the Cran archives.
Before that, you need to also install their dependancies in R:
install.packages("pheatmap")
install.packages("RJSONIO")
Then you should be able to proceed as mentioned above: install Qiimer and Biom from the command line first. Then Tax4Fun from the command line too.
If you are running on Windows you should have quite the same issues, but the installation of the different packages and dependancies is different. You can have a look at the readme at http://tax4fun.gobics.de
I have been trying to install obspy and have been running into a lot of problems. I want to install obspy which has a dependency on pyproj. But apparently obspy only works with pyproj 1.9.5.1, which I tried installing using pip (pip3 install pyproj==1.9.5.1), but only got the errors like-
_proj.c:7488:13: error: ‘PyThreadState’ {aka ‘struct _ts’} has no member named ‘exc_traceback’; did you mean ‘curexc_traceback’?
Digging deeper I found that it might be a Cython problem, and installing pyproj directly from github might help, because it would apparently make Cython recompile all the necessary files. Something along the lines of -
pip3 install git+https://github.com/jswhit/pyproj.git
However this one gives the error -
ERROR: Minimum supported proj version is 6.2.0, installed version is 5.2.0.
I di try installing a higher version of libproj-dev (sudo apt install libproj-dev=6.2.0) however it shows that there is no candidate for 6.2.0. I tried downloading the deb file and installing from that using -
sudo apt-get install ~/Downloads/libproj-dev_6.2.0-1_amd64.deb
which just leads to the error -
The following packages have unmet dependencies:
libproj-dev : Depends: libproj15 (= 6.2.0-1) but it is not installable
E: Unable to correct problems, you have held broken packages.
But I think this is not the right way to install for me anyway, since I need a specific version. Hence I tried installing directly from the tarball of the release -
pip3 install https://github.com/pyproj4/pyproj/archive/v1.9.5.1rel.tar.gz
Which leads to the first error I had, evidently due to Cython.
With errors on everything I tried to do to fix this, I am not sure what even is relevant to my problem now.
Any help is appreciated, and if this site is not the correct place for this question, please help me migrate it to its proper destination.
I am on Ubuntu 18.10.
The problem is, that Cython-generated c-files don't work for Python-3.7 if generated with Cython versions up to 0.27.3 (at least): The setup.py of pyproj (at least in the version 1.9.5.1) doesn't regenerate the_proj.c, which is generated with Cython 0.23.2 and thus the installation cannot succeed.
You have the following options:
stay on Python3.6 where everything works out of the box.
regenerate _proj.c with a current Cython-version.
For the second option:
download and unzip your prefered version from https://github.com/pyproj4/pyproj/releases/tag/v1.9.5.1rel and switch to the created folder pyproj-1.9.5.1rel.
check, that the cython-version is >=0.27.3. via cython --version.
regenerate the _proj.c file via cython -3 _proj.pyx (_proj.pyx looks like Python3-code, but also language_level=2 (i.e. cython -2 _proj.pyx) will probably work.
install running pip install .
pyproj 1.9.5.1 was release at Jan 7, 2016. At that time, the latest version Python was 3.5. In my tests. pyproj 1.9.5.1 failed to be installed on Python 3.7.4, but succeeded on Python 3.5.7.
You need to create a environment with Python 3.5 by pyenv or conda.
References
pyproj 1.9.5.1 release
Python release history
I am struggling to install packages, and "sub-packages" in Jupyter Notebook; I suspect I am missing some of the basic concepts around installing packages.
I understand that to install a package within the notebook I use
! pip install --user <package>
What I don't understand is how to install a "sub-package" (feel free to advise what the correct terminology is) such as below.
from nltk.tagger import *
Here is the original script that this comes from:
If i try :
!pip install nltk.tagger
I get the following error information / error
Collecting nltk.tagger
Could not find a version that satisfies the requirement nltk.tagger (from versions: ) No matching distribution found for nltk.tagger
So my first question is. How do I install this nltk.tagger subpackage? Also if tagger is a sub-package of NLTK, how come it isn't installed when I do a pip install NLTK?
Although the error mentions a version, searching online I can't find a reference even to the subpackage "tagger". Any advice or links explaining this would be appreciated.
if you pip inslall nltk, the subpackage nltk.tagger, and other dependencies will be installed too.
This is generally true for all packages.