Unable to install v0.11.0 tensorflow-gpu - python

I am having issues with installing the v0.11.0 version of tensorflow-gpu. My code works on tensorflow==0.11.0 but really slow. I had a tensorflow-gpu==0.11.0 but i had lost it due to an upgrade. The pip repository doesn't have a v0.11 anymore and starts with v0.12 only. I got the v0.11 wheel from here. I am able to install tensorflow==0.11.0 using the wheel at export TF_BINARY_URL=https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-0.11.0-cp27-none-linux_x86_64.whl
but I am unable to install the GPU version of tensorflow-gpu==0.11.0 using the corresponding link at export TF_BINARY_URL=https://storage.googleapis.com/tensorflow/linux/gpu/tensorflow-0.11.0-cp27-none-linux_x86_64.whl
The link looks same to me with just the folder different from gpu/ and cpu/. Upon using the link for GPU, it installs tensorflow==0.11.0 and not tensorflow-gpu=0.11.0. Where can I find a tensorflow-gpu==0.11.0 wheel?

Related

Problem in installing ecapture in python 3.10

I was just preparing to make a voice assistant and an error occurred while I was installing the ecapture module in python. I used pip for installing and the error is as shown below.
Failed to build scikit-image
ERROR: Could not build wheels for scikit-image, which is required to install py.project.toml-based projects
I have tried to install it from PyPI
even I do have tried to restart my computer, reinstall python, etc.
but it doesn't just work.
Note: only use this answer if you trust binaries built by Christoph Gohlke, who maintains an excellent index of binaries here https://www.lfd.uci.edu/~gohlke/pythonlibs/
You can either grab the needed packages from there manually, or use this package (which I wrote, full disclosure):
pip install gohlkegrabber
ggrab . scikit-image
pip install scikit_image-0.19.0-cp310-cp310-win_amd64.whl
pip install ecapture
Note that the package you were lacking is scikit-image - you may be able to find binaries elsewhere as well, the site above is only provided as a suggestion. Again, only use if you trust the author.
Also note that the package was called scikit_image-0.19.0-cp310-cp310-win_amd64.whl for me, as I'm on Python 3.10 on 64-bit Windows. Yours may have a different name (if available), but the ggrab command will tell you.
Finally note that 0.19.0 just happens to be the most recent build on that site - it's not guaranteed to have the latest build, or to have the latest build for your OS/version of Python.

Pip Errors Out When Installing Anything

I've run into an odd issue where pip just won't work. I'm using the latest version of PyCharm, and I've tried both Python 3.8.10 and 3.9.6.
I've done clean installations of both, removed from PATH, reinstalled with new installer, reinstalled pip, etc.
The error code I run into appears no matter what library I am trying to download. I've tried PyAudio, PyGame, OpenCV-Python and others.
Pip was working fine a couple days ago, and no changes have been made to any files in that time.
Attempting to run pip install --upgrade pip does not have an effect.
I've been running off virtual enviroments with setuptools 40.8.0 and pip 21.1.3 installed.
The full error for installing PyAudio can be seen here: https://pastebin.com/iQmZwsRv
PyGame error: https://pastebin.com/WEF8uM2d
OpenCV-Python Error: https://pastebin.com/LuZq99Cu
Py- audio - it's not compatible with your Py charm I suppose due to the latest install updates, try to install it on the main environment. Also, try to downgrade your version of python as many dependencies are not compatible with your version of python as stated by error[it is not compatible with this Python]
Open Cv - Fails to build the wheel package as it is attributed also in error [CMake was unable to find a build program corresponding to "Ninja". CMAKE_MAKE_PROGRAM is not set. You probably need to select a different build tool.]
May I know what build tool are you using?

Anaconda installing tensorflow and fancyimpute

As a premise, I would specify that I am new to Python, so please forgive eventual inaccuracies.
So, I have recently installed Anaconda, and updated the Python version to 3.7.1.
In order to impute some missing values in my dataset using KNN, I've found a useful function in a package called fancyimpute.
However, such package is not among those already available (that is, from Spyder, the IDE I'm using, I cannot simply import it), so I need to install it.
Opening, as such, the Anaconda prompt and typing "conda install fancyimpute" doesn't work, returning the following:
"PackagesNotFoundError: The following packages are not available from current channels:
fancyimpute
Current channels:
(here a list of some channels)
To search for alternate channels that may provide the conda package you're
looking for, navigate to
https://anaconda.org
and use the search bar at the top of the page".
Going, therefore, with "pip install fancyimpute" (although to me it's still not clear the difference between conda install and pip install) after a while I obtain the following:
"Collecting tensorflow (from fancyimpute)
Could not find a version that satisfies the requirement tensorflow (from fancyimpute) (from versions: )
No matching distribution found for tensorflow (from fancyimpute)"
I have therefore now run "conda install tensorflow", and is already about 30 minutes that is Solving environment.
What can I do? How can I obtain the desired package and avoid similar problems in future? Many thanks and hope I was clear in exposing the problem.
UPDATE: https://anaconda.org/search?q=fancyimpute From here, it seems that fancyimpute isn't available on my platform, w-64. How can I overcome this problem?
SOLVED: Apparently, I have solved the problem.
I have first created an ad hoc environment and installed tensorflow using conda.
Then, I have pip installed fancyimpute: at this point, got a couple of new errors ("Failed building wheel for fastacache", and same for cvxpy), both solved installing Microsoft Visual C++ Build Tools. So, finally, I have been able to install also fancyimpute.
Nonetheless, at this point, I coulnd't import it (ImportError: DLL load failed: The specified module could not be found. Failed to load the native TensorFlow runtime). After uninstalling and reinstalling tensorflow using conda-forge as channel, now it works.
User brittainhard on anaconda.org had the same idea. To use his/her version of the library (hosted on anaconda.org):
conda install -c brittainhard fancyimpute

Installing Cython/Statsmodels

I can not for the life of me figure out how to install the package statsmodels for Python 3. It feels like I've tried everything, but nothing is working.
Goal: import statsmodels in python
I start with:
pip install statsmodels
Main error messages:
Failed bulding wheel for statsmodels.
Cython is required to compile statsmodels from a development branch.
Please install Cython or download a source release of statsmodels.
So I try:
pip install cython
Gives:
Requirement already satisfied: cython in c:users\XXX\appdata\roaming\python\python36\site-packages
I installed cython by installing anaconda. I have tried almost all answers I could find by basic googling. But things just don't seem to be working.
Instead of listing the input\output of what I have tried I now only show the first steps, but let me know if you want to see the output of any commands.
How do I continue from here? Any help is appreciated, I've been trying to fix this for several days now and I'm only getting frustrated.
Right now I'm trying:
easy_install cython
This seemed to work, but afterwards pip install statsmodels didn't work.
Command ... failed with error code 1.
I tried:
git clone https://github.com/statsmodels/statsmodels
cd statsmodels
pip install .
And failed with a bunch of errors.
I ran into this same problem lastnight trying to install statsmodels. Finally what worked was installing the Microsoft Visual C++ compiler for Python found here: https://www.microsoft.com/en-gb/download/details.aspx?id=44266 then installing statsmodels by doing pip install statsmodels. I had already installed Cython by using pip. Though I do realize that was for Python 2.7. There doesn't seem to be this compiler available for Python 3.
The main statsmodels on PyPI is very old and uses a nonstandard setup. The release candidate for 0.10.0 is out now, and I would strongly recommend you use this one. You can install it using
pip install statsmodels=0.10.0rc2 --pre
It has wheels for all of the major platforms (Win/OSX/Linux), so you shouldn't see issues related to Cython.
To install statsmodels under anaconda, use
conda install statsmodels
The package is available in the repository of anacondata in version 0.8.0 which is the latest on pip as of now.
anaconda is not "just" a Python installation, it also contains the conda package manager that can install pre-built packages from the main repository of anaconda or from other "channels". It is not advisable to install packages that are available from the anaconda channel with pip, in general.
A comment on the issue of Cython here: to use Cython, your computer needs also a "development environment", that is the availability of a C compiler, of the Python development headers, among others. The C compiler must be compatible with the version of Python for which Cython is installed.

Problems installing TensorFlow on Mac

I am trying to follow to the installation guide on tensorflow.org and have installed Python version 2 again for that reason using Homebrew.
When I run the installation as described
$ pip install https://storage.googleapis.com/tensorflow/mac/tensorflow-0.5.0-py2-none-any.whl
I get this error message:
tensorflow-0.5.0-py2-none-any.whl is not a supported wheel on this platform.
I am obviously doing something wrong, but have no idea. Any clues?
I do not want to use virtualenv, since anaconda already comes with its own environment management conda. When installing the newest version 0.6.0 directly with pip install, I had a similar error. It seemed to not resolve the dependencies correctly.
Here is what you can try:
Install anaconda
Create a new conda workspace
Download the specific protobuf version that tensorflow needs: https://pypi.python.org/pypi/protobuf/3.0.0a3
Install it via sudo easy_install ~/Downloads/protobuf-3.0.0a3-py2.7.egg
Install a numpy version greater than 1.08.x via conda install numpy
Download the 0.6.0 version of tensorflow: https://storage.googleapis.com/tensorflow/mac/tensorflow-0.6.0-py2-none-any.whl
Install via pip install ~/Downloads/tensorflow-0.6.0-py2-none-any.whl
When you install tensorflow from the whl file directly, it should tell you when dependencies are not there. It seems not to be able to resolve these conflicts independently. My setup had issues with protobuf and numpy. After installing them manually everything worked fine.
I hope this helps!
It seems to be a common issue. Try to install it in the virtualenv. Its a much better solution, as you can always easily set up a new version of tensorflow without conflicts.
VirutalEnv Tutorial:
http://tensorflow.org/get_started/os_setup.md#virtualenv-based_installation
On the Mac, I didn't have any problem installing tensorflow with the anaconda version of python: https://www.continuum.io/downloads
The anaconda version also provides science, math, engineering, and data analysis packages. A lot of people on https://www.kaggle.com/ seem to use this...just a thought.

Categories

Resources