In [2]:

    

from collections import namedtuple
LayerParams = namedtuple(
    "LayerParams",
    [
        "percent_on_k_winner",
        "boost_strength",
        "boost_strength_factor",
        "k_inference_factor",
        "local",
        "weights_density",
    ],
    defaults=[0.25, 1.4, 0.7, 1.0, False, 0.5],
)


    

In [3]:

    

lp = LayerParams(boost_strength=0.5)


    

In [6]:

    

def prt(*args):
    print(args)
prt(*lp)


    

    




(0.25, 0.5, 0.7, 1.0, False, 0.5)

In [14]:

    

from itertools import combinations, permutations, combinations_with_replacement, product


    

In [20]:

    

for idx in  product(range(5), range(5)):
    print(idx)


    

    




(0, 0)
(0, 1)
(0, 2)
(0, 3)
(0, 4)
(1, 0)
(1, 1)
(1, 2)
(1, 3)
(1, 4)
(2, 0)
(2, 1)
(2, 2)
(2, 3)
(2, 4)
(3, 0)
(3, 1)
(3, 2)
(3, 3)
(3, 4)
(4, 0)
(4, 1)
(4, 2)
(4, 3)
(4, 4)

In [32]:

    

import torch
t = torch.rand(5,5)
q = torch.ones(5,5)


    

In [25]:

    

t[idx]


    

    
Out[25]:





tensor(0.1917)

In [30]:

    

torch.eq(t,q)


    

    
Out[30]:





tensor([[True, True, True, True, True],
        [True, True, True, True, True],
        [True, True, True, True, True],
        [True, True, True, True, True],
        [True, True, True, True, True]])

In [33]:

    

torch.allclose(t,q)


    

    
Out[33]:





False

In [34]:

    

lowest_25_hebb = [(4,0), (2,1), (0,1)]
lowest_50_mag = [(2,0), (2,2), (4,2), (0,1), (3,2), (4,1)]


    

In [40]:

    

set(lowest_25_hebb).intersection(lowest_50_mag)


    

    
Out[40]:





{(0, 1)}

In [47]:

    

corr = torch.tensor(
    [
        [0.3201, 0.8318, 0.3382, 0.9734, 0.0985],
        [0.0401, 0.8620, 0.0845, 0.3778, 0.3996],
        [0.4954, 0.0092, 0.6713, 0.8594, 0.9487],
        [0.8101, 0.0922, 0.2033, 0.7185, 0.4588],
        [0.3897, 0.6865, 0.5072, 0.9749, 0.0597],
    ]
)

weight = torch.tensor(
    [
        [19, 2, -12, 0, 0],
        [0, 0, 0, 0, 0],
        [-10, 25, -8, 0, 0],
        [21, -11, 7, 0, 0],
        [-14, 18, -6, 0, 0],
    ]
)


    

In [48]:

    

W = corr.T * weight.float()


    

In [49]:

    

W


    

    
Out[49]:





tensor([[ 6.0819,  0.0802, -5.9448,  0.0000,  0.0000],
        [ 0.0000,  0.0000,  0.0000,  0.0000,  0.0000],
        [-3.3820,  2.1125, -5.3704,  0.0000,  0.0000],
        [20.4414, -4.1558,  6.0158,  0.0000,  0.0000],
        [-1.3790,  7.1928, -5.6922,  0.0000,  0.0000]])

In [51]:

    

weights = torch.tensor([0, 10, 3, 0], dtype=torch.float)
torch.multinomial(weights, 2)


    

    
Out[51]:





tensor([2, 1])

In [58]:

    

torch.multinomial((W > 0).float(), 2)


    

    




---------------------------------------------------------------------------
RuntimeError                              Traceback (most recent call last)
<ipython-input-58-7d7f93659649> in <module>
----> 1 torch.multinomial((W > 0).float(), 2)

RuntimeError: invalid argument 2: invalid multinomial distribution (sum of probabilities <= 0) at ../aten/src/TH/generic/THTensorRandom.cpp:374

In [59]:

    

Wp = (W>0)


    

In [72]:

    

samples = torch.nonzero(Wp, as_tuple=False)


    

In [73]:

    

samples


    

    
Out[73]:





tensor([[0, 0],
        [0, 1],
        [2, 1],
        [3, 0],
        [3, 2],
        [4, 1]])

In [74]:

    

samples.shape


    

    
Out[74]:





torch.Size([6, 2])

In [75]:

    

list(samples)


    

    
Out[75]:





[tensor([0, 0]),
 tensor([0, 1]),
 tensor([2, 1]),
 tensor([3, 0]),
 tensor([3, 2]),
 tensor([4, 1])]

In [86]:

    

idx = np.random.choice(range(len(samples)), 3)
print(samples)
print(idx)
selected = samples[idx]
selected


    

    




tensor([[0, 0],
        [0, 1],
        [2, 1],
        [3, 0],
        [3, 2],
        [4, 1]])
[0 2 1]






    
Out[86]:





tensor([[0, 0],
        [2, 1],
        [0, 1]])

In [91]:

    

W[list(zip(*selected))] = True


    

In [92]:

    

W


    

    
Out[92]:





tensor([[ 1.0000,  1.0000, -5.9448,  0.0000,  0.0000],
        [ 0.0000,  0.0000,  0.0000,  0.0000,  0.0000],
        [-3.3820,  1.0000, -5.3704,  0.0000,  0.0000],
        [20.4414, -4.1558,  6.0158,  0.0000,  0.0000],
        [-1.3790,  7.1928, -5.6922,  0.0000,  0.0000]])

In [62]:

    

Wp


    

    
Out[62]:





tensor([[ True,  True, False, False, False],
        [False, False, False, False, False],
        [False,  True, False, False, False],
        [ True, False,  True, False, False],
        [False,  True, False, False, False]])

In [ ]: