+ 13

# What is the best library for working with neural network ?

What is the best library for working with deep neural network ? i worked by tensorflow and theano but their syntax is a bit hard

13 Réponses

+ 17

There's several popular libraries I know of but for python I think you should use Theano. Theano is a very flexible neural network library for use with Python. It is capable of working on CPU and GPU. I've found it to have the best documentation out of the neural net libraries I use. You can find the documentation here http://deeplearning.net/software/theano/

+ 7

try keras. it works on top of the two you mentioned, but is simpler to work with.

+ 6

Write it yourself from scratch !
Doing so will teach you a lot , and it's not too hard to do.

+ 6

"...syntax is a bit hard." It happens. If you want to look a little more at TensorFlow, just...don't forget your innovative ideas (the ones you have on your own) when you look where other people have been going:
http://playground.tensorflow.org/
- Interactive TensorFlow Visualizer (no coding)
- Adjust parameters (view node states, set functions, select features...)
- things you might use in your project, for better perspective.
https://youtu.be/4n1AHvDvVvw
- Keynote / TensorFlow Dev Summit 2017 (yesterday Feb-15; start of a day's videos)
- Code samples, release 1, where it's going, etc.
For writing your own basic bits as @Edwin Martins suggests, Coursera's Machine Learning course (Andrew Ng) walks you through some "roll-your-own" in Octave (including teaching you Octave syntax):
https://www.coursera.org/learn/machine-learning

+ 4

There's several popular libraries:
Theano - Welcome - Theano 0.7 documentation
Torch - Torch | Scientific computing for LuaJIT
Caffe - Caffe | Deep Learning Framework
TensorFlow - https://www.tensorflow.org/
MXNet - https://mxnet.readthedocs.org/en/latest/&lc=en-IN&s=1&m=968&host=www.google.co.in&ts=1486232240&sig=AJsQQ1CvlSA58ORQy-f_MPyrU2x_bCL_1g

+ 4

I found this site to be very helpfull in writing AI and simulations:
http://natureofcode.com/book/acknowledgments/

+ 3

go for keras ))

+ 3

i am using pycharm for implementing feed forwarded neural network,it is very easy to implement as you just have to import the inbuilt packages.

+ 1

several library for neural network:
-Tensor flow
opensource by google
-Torch
it is used with the Lua language
-Theano
use with Python
-Caffe
Written for C++ with CUDA
-Mxnet
source: www.quora.com/what-is-the-best-open-source-neural-network

0

I am a beginner I don't know anything about it

0

thanks my friends

0

use keras or write your own.
here is one for you.
import numpy as np
# X = (hours sleeping, hours studying), y = score on test
X = np.array(([2, 9], [1, 5], [3, 6]), dtype=float)
y = np.array(([92], [86], [89]), dtype=float)
# scale units
X = X/np.amax(X, axis=0) # maximum of X array
y = y/100 # max test score is 100
class Neural_Network(object):
def __init__(self):
#parameters
self.inputSize = 2
self.outputSize = 1
self.hiddenSize = 3
#weights
self.W1 = np.random.randn(self.inputSize, self.hiddenSize) # (3x2) weight matrix from input to hidden layer
self.W2 = np.random.randn(self.hiddenSize, self.outputSize) # (3x1) weight matrix from hidden to output layer
def forward(self, X):
#forward propagation through our network
self.z = np.dot(X, self.W1) # dot product of X (input) and first set of 3x2 weights
self.z2 = self.sigmoid(self.z) # activation function
self.z3 = np.dot(self.z2, self.W2) # dot product of hidden layer (z2) and second set of 3x1 weights
o = self.sigmoid(self.z3) # final activation function
return o
def sigmoid(self, s):
# activation function
return 1/(1+np.exp(-s))
NN = Neural_Network()
#defining our output
o = NN.forward(X)
print ("Predicted Output: \n" + str(o))
print ("Actual Output: \n" + str(y))

- 2

please teach me how to print a variable out