I need a python programmer to create a simple neural network using tensorflow and tensorboard
€30-250 EUR
Slutfört
Publicerad över sex år sedan
€30-250 EUR
Betalning vid leverans
For a smaal research project I need a python script written what uses tensorflow to create a small neural network which can is trained on this dataset: [login to view URL] and can recognise handwriting.
It should look somewhat like this:
import tensorflow as tf
from [login to view URL] import input_data
mnist = input_data.read_data_sets(one_hot = True)
n_nodes_hl1 = 500
n_nodes_hl2 = 500
n_nodes_hl3 = 500
n_classes = 10
batch_size = 100
x = [login to view URL]('float', [None, 784])
y = [login to view URL]('float')
def neural_network_model(data):
hidden_1_layer = {'weights':[login to view URL](tf.random_normal([784, n_nodes_hl1])),
'biases':[login to view URL](tf.random_normal([n_nodes_hl1]))}
hidden_2_layer = {'weights':[login to view URL](tf.random_normal([n_nodes_hl1, n_nodes_hl2])),
'biases':[login to view URL](tf.random_normal([n_nodes_hl2]))}
hidden_3_layer = {'weights':[login to view URL](tf.random_normal([n_nodes_hl2, n_nodes_hl3])),
'biases':[login to view URL](tf.random_normal([n_nodes_hl3]))}
output_layer = {'weights':[login to view URL](tf.random_normal([n_nodes_hl3, n_classes])),
'biases':[login to view URL](tf.random_normal([n_classes])),}
l1 = [login to view URL]([login to view URL](data,hidden_1_layer['weights']), hidden_1_layer['biases'])
l1 = [login to view URL](l1)
l2 = [login to view URL]([login to view URL](l1,hidden_2_layer['weights']), hidden_2_layer['biases'])
l2 = [login to view URL](l2)
l3 = [login to view URL]([login to view URL](l2,hidden_3_layer['weights']), hidden_3_layer['biases'])
l3 = [login to view URL](l3)
output = [login to view URL](l3,output_layer['weights']) + output_layer['biases']
return output
def train_neural_network(x):
prediction = neural_network_model(x)
cost = tf.reduce_mean( tf.nn.softmax_cross_entropy_with_logits(logits=prediction,labels=y) )
optimizer = [login to view URL]().minimize(cost)
hm_epochs = 10
with [login to view URL]() as sess:
[login to view URL](tf.initialize_all_variables())
for epoch in range(hm_epochs):
epoch_loss = 0
for _ in range(int(mnist.train.num_examples/batch_size)):
epoch_x, epoch_y = mnist.train.next_batch(batch_size)
_, c = [login to view URL]([optimizer, cost], feed_dict={x: epoch_x, y: epoch_y})
epoch_loss += c
print('Epoch', epoch, 'completed out of',hm_epochs,'loss:',epoch_loss)
correct = [login to view URL]([login to view URL](prediction, 1), [login to view URL](y, 1))
accuracy = tf.reduce_mean([login to view URL](correct, 'float'))
print('Accuracy:',[login to view URL]({x:[login to view URL], y:[login to view URL]}))
train_neural_network(x)
Hello sir
I am a python developer with 7 years of professional experience. I have developed many ML models with tensorflow. I am interested in this job and can start the work right now.
Best,
Zhang
€155 EUR Om 3 dagar
4,9 (127 omdömen)
7,6
7,6
5 frilansar lägger i genomsnitt anbud på €165 EUR för detta uppdrag
Hey !
I'm SAYAN PROGRAMMER
I've reviewed your complete job description, and I fulfill all the qualifications required for this project.
I have more than 15 years of experience in this field.
I am sure if you will respond me then I will be able to explain my skills as well.
I always try to provide good quality work to my clients.
My key skills are:
Coding, Neural Networks, Python, Linux, C programming , Software Architecture, PHP and much more.
Looking forward to work with you on this project.
THANKS
Regards
SAYAN PROGRAMMER