CODEWITHSUNDEEP
python
user466534
user466534

Reputation:

increment each element of list by some amount

let us suppose we have following list

w = [[[-0.82203424, -0.9185806 , 0.03494298]], [0., 0., 0.], [[1.0692896 ],[ 0.62761235],[-0.5426246 ]], [0]]

what i want is to increment each element of list by some amount using loop operator, this kind of operation we meet frequently in Neural Network Sketching, i know that there exist operation enumerate , for instance let us consider following code

import numpy as np
w = [[[-0.82203424, -0.9185806 , 0.03494298]], [0., 0., 0.], [[1.0692896 ],[ 0.62761235],[-0.5426246 ]], [0]]
for count,ele in enumerate(w): 
    print (count,ele) 
0 [[-0.82203424, -0.9185806, 0.03494298]]                                                                                       
1 [0.0, 0.0, 0.0]                                                                                                               
2 [[1.0692896], [0.62761235], [-0.5426246]]                                                                                     
3 [0]

it is clear what we can use ele arguments to access its lists, but ele has different format , there for we can also use ele to generate new enumeration right, for instance let us consider following code

import numpy as np
w = [[[-0.82203424, -0.9185806 , 0.03494298]], [0., 0., 0.], [[1.0692896 ],[ 0.62761235],[-0.5426246 ]], [0]]
for count,ele in enumerate(w): 
     for index, weight in enumerate(ele):
      print (index, weight) 
0 [-0.82203424, -0.9185806, 0.03494298]                                                                                       
0 0.0                                                                                                                         
1 0.0                                                                                                                         
2 0.0                                                                                                                         
0 [1.0692896]                                                                                                                 
1 [0.62761235]                                                                                                                
2 [-0.5426246]                                                                                                                
0 0

but how can i use this index for the original list?i have searched also and found that there exist such function numpy.ndenumerate, if i use this function in second list i will get

import numpy 
w = [[[-0.82203424, -0.9185806 , 0.03494298]], [0., 0., 0.], [[1.0692896 ],[ 0.62761235],[-0.5426246 ]], [0]]
for count,ele in enumerate(w): 
     for index, weight in numpy.ndenumerate(ele):
      print (index, weight)

result

(0, 0) -0.82203424                                                                                                            
(0, 1) -0.9185806                                                                                                             
(0, 2) 0.03494298                                                                                                             
(0,) 0.0                                                                                                                      
(1,) 0.0                                                                                                                      
(2,) 0.0                                                                                                                      
(0, 0) 1.0692896                                                                                                              
(1, 0) 0.62761235                                                                                                             
(2, 0) -0.5426246                                                                                                             
(0,) 0

so maybe accessing of original element will be following w[count][index] right?for instance accessing first element of first list will be like this

import numpy 
w = [[[-0.82203424, -0.9185806 , 0.03494298]], [0., 0., 0.], [[1.0692896 ],[ 0.62761235],[-0.5426246 ]], [0]]

for count,ele in enumerate(w): 
     for index, weight in numpy.ndenumerate(ele):
      print (index, weight) 
      print(w[count][index])

but it gives me following error 

print(w[count][index])                                                                                                    
TypeError: list indices must be integers, not tuple

how to fix? thanks in advance

Upvotes: 1

Views: 75

Answers (2)

user466534
user466534

Reputation:

here is code which is connected to neural network model and data

from google.colab import drive
drive.mount('/content/drive')
from copy import deepcopy
import numpy as np
x = [[1],[2],[3],[4]]
y = [[2],[4],[6],[8]]
w = [[[-0.82203424, -0.9185806 , 0.03494298]], [0., 0., 0.], [[1.0692896 ],[ 0.62761235],[-0.5426246 ]], [0]]
print(len(w))
def feed_forward(inputs,outputs,weights):
  pre_hidden =np.dot(inputs,weights[0])+weights[1]
  hidden =np.where(pre_hidden<0,0,pre_hidden)
  out =np.dot(hidden,weights[2])+weights[3]
  squared_error =(np.square(out-outputs))
  return squared_error
def update_weights(inputs, outputs, weights, epochs):
   for epoch in range(epochs):
      org_loss = feed_forward(inputs, outputs, weights)
      wts_tmp= deepcopy(weights)
      wts_tmp2 = deepcopy(weights)
      for i, layer in enumerate(reversed(weights)):
         for index, weight in np.ndenumerate(layer):
            wts_tmp[-(i+1)][index] += 0.0001
            loss = feed_forward(inputs, outputs, wts_tmp)
            del_loss = np.sum(org_loss -loss)/(0.0001*len(inputs))
            wts_tmp2[-(i+1)][index] += del_loss*0.01
            wts_tmp = deepcopy(weights)
      weights = deepcopy(wts_tmp2)
   return wts_tmp2
update_weights(x,y,w,1)

Upvotes: 0

Raymond Hettinger
Raymond Hettinger

Reputation: 226694

The loop should likely be:

for i, ele in enumerate(w):
    for j, x in enumerate(ele):
        print(w[i][j])           # access original using indices

The test data given is in a inconsistent format (a extra layer of square brackets on the first element):

>>> from pprint import pprint
>>> pprint(w)
[[[-0.82203424, -0.9185806, 0.03494298]],
 [0.0, 0.0, 0.0],
 [[1.0692896], [0.62761235], [-0.5426246]],
 [0]]

Perhaps this should be:

[[-0.82203424, -0.9185806, 0.03494298],
 [0.0, 0.0, 0.0],
 [1.0692896, 0.62761235, -0.5426246],
 [0]]

If the data actually does have arbitrary and inconsistent nesting, then it isn't clear what is really needed.

Upvotes: 1

Related Questions