Unable to install torch library on my linux machine.
I am trying to install sentence-transformers library in ubuntu. Got to know that Torch is a dependent library therefore I tried
pip install torch
However, at the end of installation it displays the message
Killed
I expanded my volume to cater the storage issue, however I am receiving the same error each time.
I understand very little in programming, so please answer as simply as possible. One of the notebooks on google colab throws an error. I just want to run this demo. I am attaching a screenshot, a link and the code of error.
Link to the notebook
ERROR: tensorflow 2.5.0 has requirement h5py~=3.1.0, but you'll have h5py 2.10.0 which is incompatible.
The versions are clashing.
To fix the error, run these two commands in a cell before running other cells of the notebook:
Install compatible TensorFlow version :
!pip install tensorflow==2.2.0
Install compatible h5py version
!pip install h5py==2.7.0
I am working on a machine learning project on Google Colab, it seems recently there is an issue when trying to import packages from transformers. The error message says:
ImportError: cannot import name 'SAVE_STATE_WARNING' from 'torch.optim.lr_scheduler' (/usr/local/lib/python3.7/dist-packages/torch/optim/lr_scheduler.py)
The code is simple as follow:
!pip install transformers==3.5.1
from transformers import BertTokenizer
So far I've tried to install different versions of the transformers, and import some other packages, but it seems importing any package with:
from transformers import *Package
is not working, and will result in the same error. I wonder if anyone is running into the same issue as well?
Change the torch version in colab by running this command
!pip install torch==1.4.0. Then, It worked for me.
Just change the version of tranformers to the latest one (4.5.1 at this time). That worked in colab.
!pip install transformers
The same issue occurred to me with the PyTorch version after being upgraded.
As for the solution downgrade Pytorch version to 1.4.0.
Use the below command to install
!pip install -q torch==1.4.0 -f https://download.pytorch.org/whl/cu101/torch_stable.html
It's solved a lot of problems with transformers also.
The above from udara vimukthi worked for me after trying a lot of different things, trying to get the code for "Getting started with Google BERT" to work after cloning the gitHub repository locally, so now ALL of the chapter code works while I'm showing my daughter the models.
Operating system - Windows. Running locally with GPU support, using Anaconda environment.
pip install -q --user torch==1.4.0 -f https://download.pytorch.org/whl/cu101/torch_stable.html
then I ran into some more issues and had to install the ipwidgets
pip install ipywidgets
Now it all works, as far as I've gotten. Thanks for the help with the above suggestion it saved me a lot of headaches. :)
I'm trying to install python and tensorflow to learn as part of a class I am taking, and am having some issues installing tensorflow. I keep getting the same set of errors:
C:\Users\X\AppData\Local\Programs\Python\Python36-32\python.exe: can't find '__main__' module in 'C:\\'
or
Could not find a version that satisfies the requirement tensorflow (from versions: ) No matching distribution found for tensorflow
I've tried uninstalling and reinstalling python while adding it to my path, and that solved me not being stuck with syntax errors (most of the time), but I haven't been able to make any progress. Any advice on moving forward would be appreciated.
To be clear, I am typing the following command into my command prompt:
pip3 install --upgrade tensorflow
As per the instructions on the TensorFlow website - I've been finding the "C:\>" raises a syntax error however.
your pip3 is broken, because I can clearly see tensorflow distribution on PyPi for python 3.6 https://pypi.python.org/pypi/tensorflow/1.5.0. Also, I just created a new virtual env for python 3.6 and I am able to pip install it.
Also you can download manually the "tensorflow-1.5.0-cp36-cp36m-win_amd64.whl",assuming you have windows, and pip install tensorflow-1.5.0-cp36-cp36m-win_amd64.whl
I have some imports in my jupyter notebook and among them is tensorflow:
ImportError Traceback (most recent call last)
<ipython-input-2-482704985f85> in <module>()
4 import numpy as np
5 import six.moves.copyreg as copyreg
----> 6 import tensorflow as tf
7 from six.moves import cPickle as pickle
8 from six.moves import range
ImportError: No module named tensorflow
I have it on my computer, in a special enviroment and all connected stuff also:
Requirement already satisfied (use --upgrade to upgrade): tensorflow in /Users/mac/anaconda/envs/tensorflow/lib/python2.7/site-packages
Requirement already satisfied (use --upgrade to upgrade): six>=1.10.0 in /Users/mac/anaconda/envs/tensorflow/lib/python2.7/site-packages (from tensorflow)
Requirement already satisfied (use --upgrade to upgrade): protobuf==3.0.0b2 in /Users/mac/anaconda/envs/tensorflow/lib/python2.7/site-packages (from tensorflow)
Requirement already satisfied (use --upgrade to upgrade): numpy>=1.10.1 in /Users/mac/anaconda/envs/tensorflow/lib/python2.7/site-packages (from tensorflow)
Requirement already satisfied (use --upgrade to upgrade): wheel in /Users/mac/anaconda/envs/tensorflow/lib/python2.7/site-packages (from tensorflow)
Requirement already satisfied (use --upgrade to upgrade): setuptools in ./setuptools-23.0.0-py2.7.egg (from protobuf==3.0.0b2->tensorflow)
I can import tensorflow on my computer:
>>> import tensorflow as tf
>>>
So I'm confused why this is another situation in notebook?
If you installed a TensorFlow as it said in official documentation: https://www.tensorflow.org/versions/r0.10/get_started/os_setup.html#overview
I mean creating an environment called tensorflow and tested your installation in python, but TensorFlow can not be imported in jupyter, you have to install jupyter in your tensorflow environment too:
conda install jupyter notebook
After that I run a jupyter and it can import TensorFlow too:
jupyter notebook
Jupyter runs under the conda environment where as your tensorflow install lives outside conda. In order to install tensorflow under the conda virtual environment run the following command in your terminal:
conda install -c conda-forge tensorflow
I had the same problem, and solved it by looking at the output of:
jupyter kernelspec list
which outputs the kernel information:
python2 /Users/Username/Library/Jupyter/kernels/python2
python3 /Users/Username/Library/Jupyter/kernels/python3
Notice that the path points to the Jupyter kernel for the user. To use it within the the Anaconda environment, it needs to point to the conda env you are using, and look something like Anaconda3\envs\Env_Name\share\jupyter\kernels\python3.
So, to remove the Jupyter kernelspec, just use:
jupyter kernelspec remove python3
or jupyter kernelspec remove python2 if you're using python 2
Now, the output of jupyter kernelspec list should point to the correct kernel.
See https://github.com/jupyter/notebook/issues/397 for more information about this.
Conda environment fetches the tensorflow package from the main system site-packages.
Step 1: Just deactivate conda environment
conda deactivate
pip install tensorflow
Step 2: Switch back to conda environment
conda activate YOUR_ENV_NAME
jupyter notebook
Step 3: Run the cell with import tensorflow you should be able to import.
Thanks
I also had the same problem for a long time. I wanted to import tensorflow inside the jupyter notebook within windows 10. I followed all the instructions and commands that were suggested and it was not working from the command prompt. Finally, I tried this command with the Anaconda Prompt and it worked successfully. If you are using jupyter notebook within Anaconda then go goto the windows search terminal and type "Anaconda Prompt" and inside it type following command, It will install the tensorflow inside the jupyter notebook.
conda install -c conda-forge tensorflow
the problem may when the Jupyter notebook may launching from the default but for able to import tensorflow and keras libraries so you have to install jupyter notebook like what you have installed the libraries
pip install jupyter
Run python -m ipykernel install --user --name <Environment_Name>. This should add your environment to the jupyter kernel list.
Change the kernel using Kernel->Change Kernel option or New-><Environment_Name>.
Note : Replace <Environment_Name> with the actual name of the environment.
run this command which will install tensorflow inside conda
conda install -c conda-forge tensorflow
This is what I did to fix this issue -
I installed tensorflow for windows by using below link -
https://www.tensorflow.org/install/install_windows
Once done - I activated tensorflow by using below command -
C:> activate tensorflow
(tensorflow)C:> # Your prompt should change
Once done I ran below command -
(tensorflow)C:> conda install notebook
Fetching package metadata ...........
Solving package specifications: .
Package plan for installation in environment
The following NEW packages will be INSTALLED:
bleach: 1.5.0-py35_0
colorama: 0.3.9-py35_0
decorator: 4.1.2-py35_0
entrypoints: 0.2.3-py35_0
html5lib: 0.9999999-py35_0
ipykernel: 4.6.1-py35_0
----
---
jupyter_client 100% |###############################| Time: 0:00:00 6.77 MB/s
nbformat-4.4.0 100% |###############################| Time: 0:00:00 8.10 MB/s
ipykernel-4.6. 100% |###############################| Time: 0:00:00 9.54 MB/s
nbconvert-5.2. 100% |###############################| Time: 0:00:00 9.59 MB/s
notebook-5.0.0 100% |###############################| Time: 0:00:00 8.24 MB/s
Once done I ran command
(tensorflow)C:>jupyter notebook
It opened new Juypter window and able to Run fine -
import tensorflow as tf
I was able to load tensorflow in Jupyter notebook on Windows by: first do conda create tensorflow install, then activate tensorflow at the command prompt , then execute "Jupyter notebook" from command line.
Tensorflow imports at the notebook with no error. However, I was unable to import "Pandas" &"Matplotlib, ....etc"
As suggested by #Jörg, if you have more than one kernel spec. You have to see the path it points to. In my case, it is actually the path that was to be corrected.
When I created TensorFlow virtual env, the spec had the entry for python which was pointing to base env. Thus by changing W:\\miniconda\\python.exe to W:\\miniconda\\envs\\tensorflow\\python.exe solved the problem.
So it is worth looking at your kernel spec. Delete that is not needed and keep those you want. Then look inside the JSON files where the path is given and change if needs be. I hope it helps.
There are two ways to fix this issue.
The foremost way is to create a new virtual environment and install all dependencies like jupyter notebook, tensorflow etc.
conda install jupyter notebook
conda install -c conda-forge tensorflow
The other way around is to install tensorflow in the current environment (base or any activated environment).
conda install -c conda-forge tensorflow
Note: It is advisable to create a new virtual environment for every new project. The details how to create and manage virtual environment using conda can be find here:
https://conda.io/docs/user-guide/tasks/manage-environments.html
Probably there is a problem with the TensorFlow in your environment.
In my case, After installing some libs, my TensorFlow stopped working.
So I installed TensorFlow again using pip. like so:
just run
pip install tensorflow
then I re-imported it into my jupyter notebook as :
import tensorflow as ft
In case you want to install jupyter and base libs try this:
pip install jupyter tensorflow keras numpy scipy ipython pandas matplotlib sympy nose
Other supported libraries are necessary to install with TensorFlow.Make sure if these libraries are installed:
numpy
scipy
jupyter
matplolib
pillow
scikit-learn
tensorflow-addons,
tensorflow.contrib
This worked for me. I followed this: https://www.pythonpool.com/no-module-named-tensorflow-error-solved/
TensorFlow package doesn't come by default with the root environment in Jupyter, to install it do the following :
Close Jupyter Notebook.
Open Anaconda Navigator (In windows : you can find it using the search bar)
On the sidebar, click on the Environments tab (by default you are using the root env).
You can see the installed packages, on the top switch to not-installed packages and search for tensorflow, if it doesn't show, click on Update index and it will be displayed.
The installation takes some time
If you have installed TensorFlow globally then this issue should not be occurring. As you are saying you have installed it, maybe you did it in a virtual environment.
Some background:
By default, Jupyter will open with a global python interpreter kernel.
Possible solutions:
Change your jupyter notebook kernel to your virtual environment kernel. Please check here to see how to create a kernel out of your virtual environment.
Troubleshooting:
If the above solution dint work lets do some troubleshooting. When you add your new kernel to jupyter you might have got output like below
Installed kernelspec thesis-venv in C:\Users\vishnunaik\AppData\Roaming\jupyter\kernels\venv
Check the file kernel.json in this path, which might look something like below
{
"argv": [
"C:\\Users\\vishnunaik\\Desktop\\Demo\\CodeBase\\venv\\Scripts\\python.exe",
"-m",
"ipykernel_launcher",
"-f",
"{connection_file}"
],
"display_name": "thesis-venv",
"language": "python",
"metadata": {
"debugger": true
}
}
Check the path to the python.exe is rightly pointing to your virtual environment python version or not. If not then update it accordingly.
Now you should be able to use a virtual environment in your jupyter notebook. If your kernel takes a lot of time to respond see jupyter notebook server logs, sometimes you might get output like this
[I 21:58:38.444 NotebookApp] Kernel started: adbd5551-cca3-4dad-a93f-974d7d25d53b, name: thesis-venv C:\\Users\\vishnunaik\\Desktop\\Demo\\CodeBase\\venv\\Scripts\\python.exe: No module named ipykernel_launcher
This means your virtual environment doesnot have ipykernel installed. So install it in your virtual environment using below command.
pip install ipykernel
Now you have done everything possible, so I hope this will solve your issue.