安装 turicreate 时出现 tensorflow 错误？

Question

When I install turicreate package, it gives me the following error:

Collecting tensorflow>=2.0.0 (from turicreate)
  Could not find a version that satisfies the requirement tensorflow>=2.0.0 (from turicreate) (from versions: 0.12.1, 1.0.0, 1.0.1, 1.1.0rc0, 1.1.0rc1, 1.1.0rc2, 1.1.0, 1.2.0rc0, 1.2.0rc1, 1.2.0rc2, 1.2.0, 1.2.1, 1.3.0rc0, 1.3.0rc1, 1.3.0rc2, 1.3.0, 1.4.0rc0, 1.4.0rc1, 1.4.0, 1.4.1, 1.5.0rc0, 1.5.0rc1, 1.5.0, 1.5.1, 1.6.0rc0, 1.6.0rc1, 1.6.0, 1.7.0rc0, 1.7.0rc1, 1.7.0, 1.7.1, 1.8.0rc0, 1.8.0rc1, 1.8.0, 1.9.0rc0, 1.9.0rc1, 1.9.0rc2, 1.9.0, 1.10.0rc0, 1.10.0rc1, 1.10.0, 1.10.1, 1.11.0rc0, 1.11.0rc1, 1.11.0rc2, 1.11.0, 1.12.0rc0, 1.12.0rc1, 1.12.0rc2, 1.12.0, 1.12.2, 1.12.3, 1.13.0rc0, 1.13.0rc1, 1.13.0rc2, 1.13.1, 1.13.2, 1.14.0rc0, 1.14.0rc1, 1.14.0, 2.0.0a0, 2.0.0b0, 2.0.0b1)
No matching distribution found for tensorflow>=2.0.0 (from turicreate)

which I encountered the same when installing tensorflow 2.0.0.

And I managed to install tensorflow2 with modification to the version(add a 'a0','b0','b1' after '2.0.0') using pip3 install tensorflow==2.0.0a0 . However, I still cannot pass the installation of turicreate even with tensorflow2.0.0a0 installed and result in the same 'tensorflow error' shown above. I am using python 3.6, numpy 1.16.5.

Can you please help me out?

Answer 1

Upgrade pip

# On Linux or macOS:
pip3 install -U pip

# On Windows:
python -m pip3 install -U pip

Install turicreate

pip3 install turicreate

This will solve your problem

Answer 2

I was having the same problem when I tried to install turicreate on the Jupyter Docker image that comes with some data science libraries loaded:

• Image: jupyter/scipy-notebook
• Packages: pandas, numexpr, matplotlib, scipy, seaborn, scikit-learn, scikit-image, sympy, cython, patsy, statsmodel, cloudpickle, dill, numba, bokeh, sqlalchemy, hdf5, vincent, beautifulsoup, protobuf, and xlrd packages

• Error message:

   Could not find a version that satisfies the requirement tensorflow>=2.0.0 (from turicreate) (from versions: 0.12.1, 1.0.0, 1.0.1, 1.1.0rc0, 1.1.0rc1, 1.1.0rc2, 1.1.0, 1.2.0rc0, 1.2.0rc1, 1.2.0rc2, 1.2.0, 1.2.1, 1.3.0rc0, 1.3.0rc1, 1.3.0rc2, 1.3.0, 1.4.0rc0, 1.4.0rc1, 1.4.0, 1.4.1, 1.5.0rc0, 1.5.0rc1, 1.5.0, 1.5.1, 1.6.0rc0, 1.6.0rc1, 1.6.0, 1.7.0rc0, 1.7.0rc1, 1.7.0, 1.7.1, 1.8.0rc0, 1.8.0rc1, 1.8.0, 1.9.0rc0, 1.9.0rc1, 1.9.0rc2, 1.9.0, 1.10.0rc0, 1.10.0rc1, 1.10.0, 1.10.1, 1.11.0rc0, 1.11.0rc1, 1.11.0rc2, 1.11.0, 1.12.0rc0, 1.12.0rc1, 1.12.0rc2, 1.12.0, 1.12.2, 1.12.3, 1.13.0rc0, 1.13.0rc1, 1.13.0rc2, 1.13.1, 1.13.2, 1.14.0rc0, 1.14.0rc1, 1.14.0, 2.0.0a0, 2.0.0b0, 2.0.0b1) No matching distribution found for tensorflow>=2.0.0 (from turicreate)

However, when I took the basic (stripped) Jupyter image, it worked:

• Image: jupyter/base-notebook
• Packages: git, emacs, jed, nano, tzdata, and unzip

Installation instructions for turicreate:

from __future__ import print_function
import sys

!{sys.executable} -m pip install turicreate

Output:

Successfully installed absl-py-0.9.0 astor-0.8.1 cachetools-4.0.0 coremltools-3.1 gast-0.2.2 google-auth-1.10.0 google-auth-oauthlib-0.4.1 google-pasta-0.1.8 grpcio-1.26.0 h5py-2.10.0 keras-applications-1.0.8 keras-preprocessing-1.1.0 llvmlite-0.30.0 markdown-3.1.1 numba-0.46.0 numpy-1.16.4 opt-einsum-3.1.0 pandas-0.25.3 pillow-6.2.1 prettytable-0.7.2 protobuf-3.11.2 pyasn1-0.4.8 pyasn1-modules-0.2.7 pytz-2019.3 requests-oauthlib-1.3.0 resampy-0.2.1 rsa-4.0 scipy-1.4.1 tensorboard-2.0.2 tensorflow-2.0.0 tensorflow-estimator-2.0.1 termcolor-1.1.0 turicreate-6.0 werkzeug-0.16.0 wrapt-1.11.2

So, I guess the problem is that some of the libraries that you have installed in your environment conflict with turicreate. Perhaps you should try to use the above mentioned Docker image or try to create a virtual environment so you work in isolation (not tested this).

Info on the different Docker images:

https://jupyter-docker-stacks.readthedocs.io/en/latest/using/selecting.html#jupyter-scipy-notebook

Turicreate installation instructions where they comment on the virtual environment:

https://github.com/apple/turicreate#installation

Answer 3

Firstly, give try to a latest version of pip pip3 install -U pip , usually, it has less problems than previous.

Otherwise, if above fails, you can try running with pip3 install --no-deps turicreate to get past the cryptic error. You'll need to install the dependencies manually. All 110 of them. You can generate a list of them with something along these lines:

import os;
s = os.popen('pip3 freeze turicreate').read()
for q in [x.split('==')[0] for x in s.split('\n')]:
  os.system('echo {} >> t.tmp'.format(q)

Then they can be installed with pip install -r t.tmp or for x in $(cat t.tmp); do pip3 install $x; done for x in $(cat t.tmp); do pip3 install $x; done for x in $(cat t.tmp); do pip3 install $x; done I'm not sure how the funtionality related to the tensorflow or any other failed package is affected by this but I've managed to get simple sframes working with this method. Then I've upgraded the pip and all the libs that weren't working installed without a problem.

Answer 4

Try upgrading pip. Tensorflow 2.0 needs newer versions of pip to be installed. Personally this worked,

1. upgrade pip
2. installed tf2.0
3. installed turicreate.