New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
pip install failing on wrapt uninstall #56
Comments
Thanks for this issue. wrapt is not a module requirement of DeepForest and
I've never heard of it before.
Quick googling says that this is a windows tensorflow issue upstream of
deepforest.
tensorflow/tensorflow#30191
Let's try to get tensorflow installed before DeepForest.
```
conda update
pip install tensorflow==1.14.0
```
and we'll go from there.
…On Thu, Dec 19, 2019 at 9:50 AM Joe McGlinchy ***@***.***> wrote:
Hello, I have a fresh conda environment and when running pip install
DeepForest it fails when trying to uninstall wrapt:
Cannot uninstall 'wrapt'. It is a distutils installed project and thus we
cannot accurately determine which files belong to it which would lead to
only a partial uninstall.
This is on windows 10.
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
<#56?email_source=notifications&email_token=AAJHBLDO56HS7WUYQIF7GXTQZOX75A5CNFSM4J5MDI2KYY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4IBW2DZQ>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAJHBLBKJRSC7LSWHCHE6NTQZOX75ANCNFSM4J5MDI2A>
.
--
Ben Weinstein, Ph.D.
Postdoctoral Fellow
University of Florida
http://benweinstein.weebly.com/
|
Note that DeepForest needs tensorflow < 2.0 and has been tested on tensorflow==1.14.0 |
thanks. I was able to get tensorflow installed using this suggestion tensorflow/tensorflow#30191 (comment) However, the wheel for keras-retinanet still is not able to build. |
May I see the logs?
Can you confirm how you installed keras-retinanet? There isn't a wheel. It
is right from the repo. The pip process is from the README.md
```
pip install DeepForest
pip install git+git://github.com/bw4sz/keras-retinanet.git
```
Did DeepForest install?
On Thu, Dec 19, 2019 at 10:13 AM Joe McGlinchy ***@***.***> wrote:
thanks. I was able to get tensorflow installed using this suggestion
tensorflow/tensorflow#30191 (comment)
…
However, the wheel for keras-retinanet still is not able to build.
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub, or unsubscribe.
--
Ben Weinstein, Ph.D.
Postdoctoral Fellow
University of Florida
http://benweinstein.weebly.com/
|
also, can you confirm what you did in the other issue, i'm making a windows install note in the docs: https://deepforest.readthedocs.io/en/latest/installation.html#windows-installation |
here's what I did:
|
Couple quick things. But I'm going to open up an azure windows instance to try to replicate.
will give you 2.0, but < 2.0 is required by DeepForest, which is why you can see that pip install DeepForest uninstalls the tensorflow you just installed, and then installs tensorflow 1.14.0. I'll test making any < 2.0 available, but for the moment
is preferred. For retinanet, I can see that you actually install fine, its building the C library after install which fails. That message means you do not have a valid c++ compiler on path. https://stackoverflow.com/questions/41724445/python-pip-on-windows-command-cl-exe-failed After we finish i'll try to brainstorm ways to wrap this together, maybe with pyinstaller for windows. |
I don't have a ton of experience here, i switch to mac for exactly this
reason. But my strategy used to be to use the visual studio build tools and
add them to path
https://visualstudio.microsoft.com/downloads/#build-tools-for-visual-studio-2017
you don't need the whole visual studio (so big), just the tiny build tools.
I'm looking into a windows build to create wheels, or conda forge
…On Fri, Dec 20, 2019 at 8:39 AM Ben Weinstein ***@***.***> wrote:
Couple quick things. But I'm going to open up an azure windows instance to
try to replicate.
pip install tensorflow
will give you 2.0, but < 2.0 is required by DeepForest, which is why you
can see that pip install DeepForest uninstalls the tensorflow you just
installed, and then installs tensorflow 1.14.0. I'll test making any < 2.0
available, but for the moment
pip install tensorflow==1.14.0
is preferred.
For retinanet, I can see that you actually install file, its building the
C library after install which fails. You do not have a valid c++ compiler
on path.
https://stackoverflow.com/questions/41724445/python-pip-on-windows-command-cl-exe-failed
After we finish i'll try to brainstorm ways to wrap this together, maybe
with pyinstaller for windows.
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
<#56?email_source=notifications&email_token=AAJHBLFIQO4IKLWIXDBRP7TQZTYLXA5CNFSM4J5MDI2KYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEHNNASY#issuecomment-567988299>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAJHBLAUOKXHSAIDQVEYF73QZTYLXANCNFSM4J5MDI2A>
.
--
Ben Weinstein, Ph.D.
Postdoctoral Fellow
University of Florida
http://benweinstein.weebly.com/
|
This might be sufficient, https://wiki.python.org/moin/WindowsCompilers will test on azure |
@joemcglinchy , I'm building a separate wheel for windows for the compiled c++ code. I will update here. |
#60 new compiled wheels for many python version and operating systems, this should make install much easier. On a new azure windows 2019 data science image (https://azure.microsoft.com/en-us/services/virtual-machines/data-science-virtual-machines/)
Check the code on the front of the repo
Looks good to me. Closing. Feel free to open if new install fails. |
Hello, I have a fresh conda environment and when running
pip install DeepForest
it fails when trying to uninstallwrapt
:Cannot uninstall 'wrapt'. It is a distutils installed project and thus we cannot accurately determine which files belong to it which would lead to only a partial uninstall.
This is on windows 10.
The text was updated successfully, but these errors were encountered: