Comments (44)
I added '-DCMAKE_GENERATOR_PLATFORM=x64' flag to the setup.py and the problem solved in Win64 (Win10) and Python 3.5.4.
if 0 != execute(['cmake',
'-DCMAKE_GENERATOR_PLATFORM=x64', <----- THIS
'-DCMAKE_BUILD_TYPE={}'.format(build_type),
'-DCMAKE_VERBOSE_MAKEFILE={}'.format(int(self.verbose)),
from multicore-tsne.
The above solution works for me. Let me summarize the solution:
- If you installed package before,
pip uninstall MulticoreTSNE
- Clone the source
git clone https://github.com/DmitryUlyanov/Multicore-TSNE.git
- Edit
/Multicore-TSNE/setup.py
,
Find these piece of code and add one line
if 0 != execute(['cmake',
'-DCMAKE_GENERATOR_PLATFORM=x64', <----- THIS
'-DCMAKE_BUILD_TYPE={}'.format(build_type),
'-DCMAKE_VERBOSE_MAKEFILE={}'.format(int(self.verbose)),
- install by source
cd /Multicore-TSNE
pip install .
from multicore-tsne.
The problem I am having is that the file tsne_multicore.dll has been compiled with i386 architecture (32-bit) so it will not work on 64-bit systems, thus the "error 0xc1". I am not sure if this is the correct answer for everyone, but I was able to add:
if 0 != execute(['cmake',
'-GMinGW Makefiles', <----------This
'-DCMAKE_BUILD_TYPE={}'.format(build_type),
'-DCMAKE_VERBOSE_MAKEFILE={}'.format(int(self.verbose)),
to the setup.py file to get it to compile with MinGW for WIndows which works and produces the right files for the architecture.
from multicore-tsne.
I got the same problem with python 3.4 when running the example inside the installation directory "Multicore-TSNE"
Worked if I run python outside the installation directory.
from multicore-tsne.
I added '-DCMAKE_GENERATOR_PLATFORM=x64' flag to the setup.py and the problem solved in Win64 (Win10) and Python 3.5.4.
if 0 != execute(['cmake',
'-DCMAKE_GENERATOR_PLATFORM=x64', <----- THIS
'-DCMAKE_BUILD_TYPE={}'.format(build_type),
'-DCMAKE_VERBOSE_MAKEFILE={}'.format(int(self.verbose)),Thanx this solved my problem on win10 & python 3.5.4
This works for me as well. If you had already installed the package, just type pip uninstall MulticoreTSNE and then reinstall it..
from multicore-tsne.
I had the same problem. This is how I fixed it:
- go to the parent directory
- run
pip install Multicore-TSNE/
then the egg is added to/usr/local/lib/python2.7/dist-packages/
. Same can be done on python3.x with pip3.
from multicore-tsne.
I have same problem on Win 10, tried to install using all methods mentioned.
from multicore-tsne.
@jroakes THX! That worked perfectly!
from multicore-tsne.
I added '-DCMAKE_GENERATOR_PLATFORM=x64' flag to the setup.py and the problem solved in Win64 (Win10) and Python 3.5.4.
if 0 != execute(['cmake',
'-DCMAKE_GENERATOR_PLATFORM=x64', <----- THIS
'-DCMAKE_BUILD_TYPE={}'.format(build_type),
'-DCMAKE_VERBOSE_MAKEFILE={}'.format(int(self.verbose)),
Thanx this solved my problem on win10 & python 3.5.4
from multicore-tsne.
I just tried to replicate this and it works fine for me (python 3.6.2).
Which python version are you running?
from multicore-tsne.
from multicore-tsne.
which command did you use to install?
from multicore-tsne.
git clone https://github.com/DmitryUlyanov/Multicore-TSNE.git
cd Multicore-TSNE/
pip install --no-cache-dir .
from multicore-tsne.
try without --no-cache-dir
from multicore-tsne.
from multicore-tsne.
I am having the same problem on 3.6.2.
from multicore-tsne.
same problem
from multicore-tsne.
any updates regarding this? i have the same issue.
from multicore-tsne.
Please try to install with python setup.py install
from multicore-tsne.
I solved the problem by compiling the shared .so file manually using g++
from multicore-tsne.
Python 2.7
Compiled using windows build tools 2017
Installed via pip install. and python setup.py install
Neither work.
Any idea why?
from multicore-tsne.
Tried both python setup.py install
, running pip install .
and running pip install Multicore-TSNE
from the parent dir, neither works on Windows 10 with 2017 devtools.
from multicore-tsne.
I was experiencing this on one of my machines and so I just updated numpy in my Anaconda distribution. That seemed to fix it.
from multicore-tsne.
@DmitryUlyanov could you kindly take a look, there is still a problem on Windows...
UPD
After server reboot somehow it started working.. )
from multicore-tsne.
Hi, I also have the same problem on Win 10, and also tried with all the methods mentioned and couldn't make it to work.
from multicore-tsne.
I also have the same issue on Win 10.
from multicore-tsne.
Same issue on Win 10. Tried everything above.
from multicore-tsne.
Did anyone find any solution? I am getting the same error
from multicore-tsne.
If you are on Windows,
- Install MinGW 64. I had it already, but you can install from http://mingw-w64.org/doku.php.
- Then add MinGW flag as #32 (comment) describes.
- Next, build with
python setup.py install
- and install with
pip install .
from multicore-tsne.
I spent an hour trying different installs with cmd, mingw, pip, setup.py. I'm using python 2.7.10 on windows7. Nothing worked.
I then tried on python3.6, worked immediately.
from multicore-tsne.
I built the lib using MSVS 2017 and it worked. I can share the file. But there is an easier way: use pypi wheel package (pip regular installation process)! It simply works!
from multicore-tsne.
I added '-DCMAKE_GENERATOR_PLATFORM=x64' flag to the setup.py and the problem solved in Win64 (Win10) and Python 3.5.4.
if 0 != execute(['cmake',
'-DCMAKE_GENERATOR_PLATFORM=x64', <----- THIS
'-DCMAKE_BUILD_TYPE={}'.format(build_type),
'-DCMAKE_VERBOSE_MAKEFILE={}'.format(int(self.verbose)),
thx, solved my problem. win 10
from multicore-tsne.
Hi, I used pip install MulticoreTSNE and got the same error while running my code. Anyone got the solution to fix it?
thanks!
from multicore-tsne.
I added '-DCMAKE_GENERATOR_PLATFORM=x64' flag to the setup.py and the problem solved in Win64 (Win10) and Python 3.5.4.
if 0 != execute(['cmake',
'-DCMAKE_GENERATOR_PLATFORM=x64', <----- THIS
'-DCMAKE_BUILD_TYPE={}'.format(build_type),
'-DCMAKE_VERBOSE_MAKEFILE={}'.format(int(self.verbose)),thx, solved my problem. win 10
where I can find the setup.py file, please?
from multicore-tsne.
If you are on Windows,
- Install MinGW 64. I had it already, but you can install from http://mingw-w64.org/doku.php.
- Then add MinGW flag as #32 (comment) describes.
- Next, build with
python setup.py install
- and install with
pip install .
hi
can you help me , what do you mean by "build with python setup.py install" then install with
pip install .` ?
can you provide me with full commands please
from multicore-tsne.
If you are on Windows,
- Install MinGW 64. I had it already, but you can install from http://mingw-w64.org/doku.php.
- Then add MinGW flag as #32 (comment) describes.
- Next, build with
python setup.py install
- and install with
pip install .
hi
can you help me , what do you mean by "build withpython setup.py install" then install with
pip install .` ?
can you provide me with full commands please
I alrdy forgot what i did to solve the problem. But i think the steps are:
- download package from here https://github.com/DmitryUlyanov/Multicore-TSNE
- open setup.py install with notepad or jupyter
- change the code as was written above
- install the pakage by command in cmd. python setup.py install
from multicore-tsne.
Also having this issue on OSX, @akupemula, can you expand on how to compile the .so with g++ manually? Thanks!
from multicore-tsne.
i've done everything mentioned here #32 (comment)
compiled with cmake x64 and mingw-w64\x86_64-8.1.0-posix-sjlj-rt_v6-rev0
but still getting
"RuntimeError: Cannot find/open tsne_multicore shared library"
from multicore-tsne.
from multicore-tsne.
I added '-DCMAKE_GENERATOR_PLATFORM=x64' flag to the setup.py and the problem solved in Win64 (Win10) and Python 3.5.4.
if 0 != execute(['cmake',
'-DCMAKE_GENERATOR_PLATFORM=x64', <----- THIS
'-DCMAKE_BUILD_TYPE={}'.format(build_type),
'-DCMAKE_VERBOSE_MAKEFILE={}'.format(int(self.verbose)),
this helped me on win 10 x64 python 3.6.8
from multicore-tsne.
I added '-DCMAKE_GENERATOR_PLATFORM=x64' flag to the setup.py and the problem solved in Win64 (Win10) and Python 3.5.4.
if 0 != execute(['cmake',
'-DCMAKE_GENERATOR_PLATFORM=x64', <----- THIS
'-DCMAKE_BUILD_TYPE={}'.format(build_type),
'-DCMAKE_VERBOSE_MAKEFILE={}'.format(int(self.verbose)),
Thank you very much, this solved my problem,this helped me on windows 10 x64 python 3.6.5
from multicore-tsne.
Does anyone know hot to fix it on Mac OS?
from multicore-tsne.
Problem:
demo:
from sklearn.datasets import load_digits
from MulticoreTSNE import MulticoreTSNE as TSNE
from matplotlib import pyplot as plt
digits = load_digits()
embeddings = TSNE(n_jobs=4).fit_transform(digits.data)
vis_x = embeddings[:, 0]
vis_y = embeddings[:, 1]
plt.scatter(vis_x, vis_y, c=digits.target, cmap=plt.cm.get_cmap("jet", 10), marker='.')
plt.colorbar(ticks=range(10))
plt.clim(-0.5, 9.5)
plt.show()
If you run this demo, the error may occurr:
Traceback (most recent call last):
File "/xxxx/miniconda3/xxx/test_MultiCore_TSNE.py", line 10, in <module>
embeddings = TSNE(n_jobs=4).fit_transform(digits.data)
File "/xxxx/miniconda3/xxx/site-packages/MulticoreTSNE/__init__.py", line 103, in __init__
raise RuntimeError('Cannot find/open tsne_multicore shared library')
RuntimeError: Cannot find/open tsne_multicore shared library
Explore:
Open /xxx//miniconda3/xxx/site-packages/MulticoreTSNE/__init__.py
to check what makes it happen,and locate to around line 103:
path = os.path.dirname(os.path.realpath(__file__))
try:
sofile = (glob(os.path.join(path, "libtsne*.so")) + glob(os.path.join(path, "*tsne*.dll")))[0]
self.C = self.ffi.dlopen(os.path.join(path, sofile))
except (IndexError, OSError):
raise RuntimeError('Cannot find/open tsne_multicore shared library')
To get more concrete information about this error, we can modified this source code as follow:
path = os.path.dirname(os.path.realpath(__file__))
try:
sofile = (glob(os.path.join(path, "libtsne*.so")) + glob(os.path.join(path, "*tsne*.dll")))[0]
self.C = self.ffi.dlopen(os.path.join(path, sofile))
# except (IndexError, OSError):
# raise RuntimeError('Cannot find/open tsne_multicore shared library')
except IndexError as e:
print(e)
exit(-1)
except OSError as e1:
print(e1)
exit(-1)
Then run the demo again, errors may occurr as follow:
cannot load library '/xxxx/miniconda3/xxx/site-packages/MulticoreTSNE/libtsne_multicore.so':
/xxxx/miniconda3/xxx/site-packages/scipy/_lib/_uarray/../../../../../libstdc++.so.6: version `GLIBCXX_3.4.29' not found
(required by /xxxx/miniconda3/xxx/site-packages/MulticoreTSNE/libtsne_multicore.so)
From this information, we can see that there's no GLIBCXX_3.4.29
in libstdc++.so.6
. Test it as follow:
$ real path /xxxxx/miniconda3/xxx/scipy/_lib/_uarray/../../../../../libstdc++.so.6
$ /xxxx/miniconda3/xxx/lib/libstdc++.so.6.0.26
$ strings /xxxx/miniconda3/xxx/lib/libstdc++.so.6.0.26 | grep GLIBCXX_3.4.29
$ # return nothing
Solution
Reference:https://blog.csdn.net/qq_39295044/article/details/86685729
We can find /usr/lib/libstdc++.so.6.0.29
contains GLIBCXX_3.4.29
, test it:
$ strings /usr/lib/libstdc++.so.6 | grep GLIBCXX_3.4.29
GLIBCXX_3.4.29
GLIBCXX_3.4.29
This lib file satisfies the requirement, so we can replace /xxxx/miniconda3/xxx/lib/libstdc++.so.6.0.26
with /usr/lib/libstdc++.so.6.0.29
:
$ mv /xxxx/miniconda3/xxx/lib/libstdc++.so.6.0.26 /xxxx/miniconda3/xxx/lib/libstdc++.so.6.0.26_bak
$ ln -s /usr/lib/libstdc++.so.6.0.29 /xxxx/miniconda3/xxx/lib/libstdc++.so.6.0.26
Run the demo again, good lucky!
from multicore-tsne.
I also encountered a similar issue, but I managed to resolve it by following these steps:
-
First, I downloaded the Mingw64 compiler for Windows from the official website: https://winlibs.com/#download-release.
-
Once the download was complete, I installed the Mingw64 compiler on my system.
-
After installing Mingw64, I changed the following code to build the project:
if 0 != execute(['cmake',
'-G', 'MinGW Makefiles', <----- This Line
'-DCMAKE_BUILD_TYPE={}'.format(build_type),
- I compiled the project using the Mingw64 compiler, and fortunately, it worked perfectly fine.
By following these steps and utilizing the Mingw64 compiler, I was able to overcome the problem.
MulticoreTSNE-0.1-cp37-cp37m-win_amd64.zip
from multicore-tsne.
Related Issues (20)
- Memory Allocation Fail (Big Data)
- TSNE results in dense centred ball HOT 2
- Ruby Library
- License Clarification HOT 1
- Failed building wheel for MulticoreTSNE HOT 1
- Non-verbose crash on input containing NaN
- Error in builds using cmake >= 3.20? HOT 8
- Cannot generate Makefile. HOT 1
- Can't install on Apple M1 chip HOT 13
- Enabling verbose causes kernel crash (likely a divided by zero)
- `pip install cmake==3.18.4` worked for me.
- cmake version requirement.txt HOT 1
- signature of sklearn.datasets.make_blobs has changed
- Fix compiling HOT 5
- why can't Multicore-TSNE speed up my experiment with mnist
- linux compile
- Error in generate MING64 Makefile
- Type error with sklearn 1.2.3 HOT 1
- Static library is built but runtime is trying to load dynamic one HOT 1
- Graphics problem with tsne algorithm
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from multicore-tsne.