analitics

Pages

Showing posts with label Theano. Show all posts
Showing posts with label Theano. Show all posts

Wednesday, September 11, 2019

Python 3.7.4 : Using the theano pakage.

If you want to test theano then you need to see this webpage.
[root@desk mythcat]# dnf search theano
======================== Name & Summary Matched: theano ========================
python-theano-doc.noarch : Theano documentation
============================= Name Matched: theano =============================
python3-theano.noarch : Mathematical expressions involving multidimensional
                      : arrays
=========================== Summary Matched: theano ============================
python3-lasagne.noarch : Lightweight library to build and train neural networks
                       : in Theano
[root@desk mythcat]# pip3 install Theano --user
WARNING: Running pip install with root privileges is generally not a good idea. 
Try `pip3 install --user` instead.
Collecting Theano
...
  Running setup.py install for Theano ... done
Successfully installed Theano-1.0.4 scipy-1.3.1
Let's see first example:
[mythcat@desk ~]$ python3 
Python 3.7.4 (default, Jul  9 2019, 16:32:37) 
[GCC 9.1.1 20190503 (Red Hat 9.1.1-1)] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import theano
/home/mythcat/.local/lib/python3.7/site-packages/theano/configdefaults.py:560: 
UserWarning: DeprecationWarning: there is no c++ compiler.This is deprecated and with 
Theano 0.11 a c++ compiler will be mandatory
  warnings.warn("DeprecationWarning: there is no c++ compiler."
WARNING (theano.configdefaults): g++ not detected ! Theano will be unable to execute 
optimized C-implementations (for both CPU and GPU) and will default to Python implementations.
 Performance will be severely degraded. To remove this warning, set Theano flags cxx to an empty
 string.
WARNING (theano.tensor.blas): Using NumPy C-API based implementation for BLAS functions.
>>> import theano
>>> import theano.tensor as T
>>> x = T.dmatrix('x')
>>> s = 1 / (1 + T.exp(-x))
>>> logistic = theano.function([x], s)
>>> logistic([[0, 1], [-1, -2]])
array([[0.5       , 0.73105858],
       [0.26894142, 0.11920292]])
>>> ... 

Sunday, December 18, 2016

NVIDIA python module Theano.

I try to use python module Theano.
First I install this python module.
C:\WINDOWS\system32>cd C:\Python27

C:\Python27>cd Scripts

C:\Python27\Scripts>pip install Theano
Collecting Theano
Using cached Theano-0.8.2.tar.gz
Requirement already satisfied: numpy>=1.7.1 in c:\python27\lib\site-packages (from Theano)
Requirement already satisfied: scipy>=0.11 in c:\python27\lib\site-packages (from Theano)
Requirement already satisfied: six>=1.9.0 in c:\python27\lib\site-packages (from Theano)
Installing collected packages: Theano
Running setup.py install for Theano ... done
Successfully installed Theano-0.8.2

When I try to used I got this error:
import theano
WARNING (theano.configdefaults): g++ not detected ! Theano will be unable to execute optimized C-implementations (for both CPU and GPU) and will default to Python implementations. Performance will be severely degraded. To remove this warning, set Theano flags cxx to an empty string.

I try to fix that error, but I don't find any solution.
This python module work. I tested with examples from NVIDIA, see: