Matrix-vector-multiplication with tensors in numpy - python

I have a numpy.array A with shape (l,l) and another numpy.array B with shape (l,m,n). Usually, the second and third dimension in B correspond to spatial cells and the first to something else.
I want to compute
l,m,n = 2,3,4 # dummy dimensions
A = np.random.rand(l,l) # dummy data
B = np.random.rand(l,m,n) # dummy data
C = np.zeros((l,m,n))
for i in range(m):
for j in range(n):
C[:,i,j] = A#B[:,i,j]
i.e., in every spatial cell, I want to perform a matrix-vector-multiplication.
Since I have to do this frequently, I would like to know, if there's a more compact way to write this with numpy. (Especially, because there are several situations in which the tensor has shape (l,m,n,o,p).)
Thank you in advance!

I found the answer using np.einsum:
np.einsum('ij,jkl->ikl', A,B)
Explanation:
Einstein notation implies that we sum over matching subscripts.
np.einsum('ij,jkl->ikl', A,B)
= rewritten in math terms
A_{i,j} B_{j,k,l}
= Einstein notation implies summation
sum_j A_{i,j} B_{j,k,l}

Related

1D Convolution of 2D arrays

I have 2 arrays of sets of signals, both 16x90000 arrays. In other words, 2 arrays with 16 signals in each. I want to perform matched filtering on the signals, row by row, correlating row 1 of array 1 with row 1 of array 2, and so forth. I've tried using scipy's signal.convolve2D but it is extremely slow, taking tens of seconds to convolve even a 2x90000 array. I'm not sure if I am simply implementing wrong, or if there is a more efficient way of achieving what I want. I know the arrays are long, but I feel it should still be achievable. I have a feeling convolve2d is actually convolving to a squared factor higher than I want and convolving rows by columns too but I may be misunderstanding.
My implementation:
A.shape = (16,90000) # an array of 16 signals each 90000 samples long
B.shape = (16,90000) # another array of 16 signals each 90000 samples long
corr = sig.convolve2d(A,B,mode='same')
I haven't had much coffee yet so there's every chance I'm being stupid right now.
Please no for loops.
Since you need to correlate the signals row by row, the most basic solution would be:
import numpy as np
from scipy.signal import correlate
# sample inputs: A and B both have n signals of length m
n, m = 2, 5
A = np.random.randn(n, m)
B = np.random.randn(n, m)
C = np.vstack([correlate(a, b, mode="same") for a, b in zip(A, B)])
# [[-0.98455996 0.86994062 -1.1446486 -2.1751074 -0.59270322]
# [ 1.7945015 1.51317292 1.74286042 -0.57750712 -1.9178488 ]]]
One way to avoid a looped solution could be by bootlegging off a deep learning library, like PyTorch. Torch's Conv1d (though named conv, it effectively performs cross-correlation) can handle this scenario.
import torch
import torch.nn.functional as F
# Convert A and B to torch tensors
P = torch.from_numpy(A).unsqueeze(0) # (1, n, m)
Q = torch.from_numpy(B).unsqueeze(1) # (n, 1, m)
# Use conv1d --- with groups = n
def torch_correlate(A, B, n):
with torch.no_grad():
return F.conv1d(A, B, bias=None, stride=1, groups=n, padding="same").squeeze(0).numpy()
R = torch_correlate(P, Q, n)
# [[-0.98455996 0.86994062 -1.1446486 -2.1751074 -0.59270322]
# [ 1.7945015 1.51317292 1.74286042 -0.57750712 -1.9178488 ]]
However, I believe there shouldn't be any significant difference in the results, since grouping might be using some form of iteration internally as well. (Plus there is an overhead of converting from torch to numpy and back to consider).
I would suggest using the first method generally. Unless if you are working on really large signals, then you could theoretically use the PyTorch version to run it really fast on GPU, which you won't be able to do with the regular scipy one.

Python - matrix multiplication

i have an array y with shape (n,), I want to compute the inner product matrix, which is a n * n matrix
However, when I tried to do it in Python
np.dot(y , y)
I got the answer n, this is not what I am looking for
I have also tried:
np.dot(np.transpose(y),y)
np.dot(y, np.transpose(y))
I always get the same answer n
I think you are looking for:
np.multiply.outer(y,y)
or equally:
y = y[None,:]
y.T#y
example:
y = np.array([1,2,3])[None,:]
output:
#[[1 2 3]
# [2 4 6]
# [3 6 9]]
You can try to reshape y from shape (70,) to (70,1) before multiplying the 2 matrices.
# Reshape
y = y.reshape(70,1)
# Either below code would work
y*y.T
np.matmul(y,y.T)
One-liner?
np.dot(a[:, None], a[None, :])
transpose doesn't work on 1-D arrays, because you need atleast two axes to 'swap' them. This solution adds a new axis to the array; in the first argument, it looks like a column vector and has two axes; in the second argument it still looks like a row vector but has two axes.
Looks like what you need is the # matrix multiplication operator. dot method is only to compute dot product between vectors, what you want is matrix multiplication.
>>> a = np.random.rand(70, 1)
>>> (a # a.T).shape
(70, 70)
UPDATE:
Above answer is incorrect. dot does the same things if the array is 2D. See the docs here.
np.dot computes the dot product of two arrays. Specifically,
If both a and b are 1-D arrays, it is inner product of vectors (without complex conjugation).
If both a and b are 2-D arrays, it is matrix multiplication, but using matmul or a # b is preferred.
Simplest way to do what you want is to convert the vector to a matrix first using np.matrix and then using the #. Although, dot can also be used # is better because conventionally dot is used for vectors and # for matrices.
>>> a = np.random.rand(70)
(70,)
>>> a.shape
>>> a = np.matrix(a).T
>>> a.shape
(70, 1)
>>> (a # a.T).shape
(70, 70)

np.bincount for 1 line, vectorized multidimensional averaging

I am trying to vectorize an operation using numpy, which I use in a python script that I have profiled, and found this operation to be the bottleneck and so needs to be optimized since I will run it many times.
The operation is on a data set of two parts. First, a large set (n) of 1D vectors of different lengths (with maximum length, Lmax) whose elements are integers from 1 to maxvalue. The set of vectors is arranged in a 2D array, data, of size (num_samples,Lmax) with trailing elements in each row zeroed. The second part is a set of scalar floats, one associated with each vector, that I have a computed and which depend on its length and the integer-value at each position. The set of scalars is made into a 1D array, Y, of size num_samples.
The desired operation is to form the average of Y over the n samples, as a function of (value,position along length,length).
This entire operation can be vectorized in matlab with use of the accumarray function: by using 3 2D arrays of the same size as data, whose elements are the corresponding value, position, and length indices of the desired final array:
sz_Y = num_samples;
sz_len = Lmax
sz_pos = Lmax
sz_val = maxvalue
ind_len = repmat( 1:sz_len ,1 ,sz_samples);
ind_pos = repmat( 1:sz_pos ,sz_samples,1 );
ind_val = data
ind_Y = repmat((1:sz_Y)',1 ,Lmax );
copiedY=Y(ind_Y);
mask = data>0;
finalarr=accumarray({ind_val(mask),ind_pos(mask),ind_len(mask)},copiedY(mask), [sz_val sz_pos sz_len])/sz_val;
I was hoping to emulate this implementation with np.bincounts. However, np.bincounts differs to accumarray in two relevant ways:
both arguments must be of same 1D size, and
there is no option to choose the shape of the output array.
In the above usage of accumarray, the list of indices, {ind_val(mask),ind_pos(mask),ind_len(mask)}, is 1D cell array of 1x3 arrays used as index tuples, while in np.bincounts it must be 1D scalars as far as I understand. I expect np.ravel may be useful but am not sure how to use it here to do what I want. I am coming to python from matlab and some things do not translate directly, e.g. the colon operator which ravels in opposite order to ravel. So my question is how might I use np.bincount or any other numpy method to achieve an efficient python implementation of this operation.
EDIT: To avoid wasting time: for these multiD index problems with complicated index manipulation, is the recommend route to just use cython to implement the loops explicity?
EDIT2: Alternative Python implementation I just came up with.
Here is a heavy ram solution:
First precalculate:
Using index units for length (i.e., length 1 =0) make a 4D bool array, size (num_samples,Lmax+1,Lmax+1,maxvalue) , holding where the conditions are satisfied for each value in Y.
ALLcond=np.zeros((num_samples,Lmax+1,Lmax+1,maxvalue+1),dtype='bool')
for l in range(Lmax+1):
for i in range(Lmax+1):
for v in range(maxvalue+!):
ALLcond[:,l,i,v]=(data[:,i]==v) & (Lvec==l)`
Where Lvec=[len(row) for row in data]. Then get the indices for these using np.where and initialize a 4D float array into which you will assign the values of Y:
[indY,ind_len,ind_pos,ind_val]=np.where(ALLcond)
Yval=np.zeros(np.shape(ALLcond),dtype='float')
Now in the loop in which I have to perform the operation, I compute it with the two lines:
Yval[ind_Y,ind_len,ind_pos,ind_val]=Y[ind_Y]
Y_avg=sum(Yval)/num_samples
This gives a factor of 4 or so speed up over the direct loop implementation. I was expecting more. Perhaps, this is a more tangible implementation for Python heads to digest. Any faster suggestions are welcome :)
One way is to convert the 3 "indices" to a linear index and then apply bincount. Numpy's ravel_multi_index is essentially the same as MATLAB's sub2ind. So the ported code could be something like:
shape = (Lmax+1, Lmax+1, maxvalue+1)
posvec = np.arange(1, Lmax+1)
ind_len = np.tile(Lvec[:,None], [1, Lmax])
ind_pos = np.tile(posvec, [n, 1])
ind_val = data
Y_copied = np.tile(Y[:,None], [1, Lmax])
mask = posvec <= Lvec[:,None] # fill-value independent
lin_idx = np.ravel_multi_index((ind_len[mask], ind_pos[mask], ind_val[mask]), shape)
Y_avg = np.bincount(lin_idx, weights=Y_copied[mask], minlength=np.prod(shape)) / n
Y_avg.shape = shape
This is assuming data has shape (n, Lmax), Lvec is Numpy array, etc. You may need to adapt the code a little to get rid of off-by-one errors.
One could argue that the tile operations are not very efficient and not very "numpythonic". Something with broadcast_arrays could be nice, but I think I prefer this way:
shape = (Lmax+1, Lmax+1, maxvalue+1)
posvec = np.arange(1, Lmax+1)
len_idx = np.repeat(Lvec, Lvec)
pos_idx = np.broadcast_to(posvec, data.shape)[mask]
val_idx = data[mask]
Y_copied = np.repeat(Y, Lvec)
mask = posvec <= Lvec[:,None] # fill-value independent
lin_idx = np.ravel_multi_index((len_idx, pos_idx, val_idx), shape)
Y_avg = np.bincount(lin_idx, weights=Y_copied, minlength=np.prod(shape)) / n
Y_avg.shape = shape
Note broadcast_to was added in Numpy 1.10.0.

How to vectorize 3D Numpy arrays

I have a 3D numpy array like a = np.zeros((100,100, 20)). I want to perform an operation over every x,y position that involves all the elements over the z axis and the result is stored in an array like b = np.zeros((100,100)) on the same corresponding x,y position.
Now i'm doing it using a for loop:
d_n = np.array([...]) # a parameter with the same shape as b
for (x,y), v in np.ndenumerate(b):
C = a[x,y,:]
### calculate some_value using C
minv = sys.maxint
depth = -1
C = a[x,y,:]
for d in range(len(C)):
e = 2.5 * float(math.pow(d_n[x,y] - d, 2)) + C[d] * 0.05
if e < minv:
minv = e
depth = d
some_value = depth
if depth == -1:
some_value = len(C) - 1
###
b[x,y] = some_value
The problem now is that this operation is much slower than others done the pythonic way, e.g. c = b * b (I actually profiled this function and it's around 2 orders of magnitude slower than others using numpy built in functions and vectorized functions, over a similar number of elements)
How can I improve the performance of such kind of functions mapping a 3D array to a 2D one?
What is usually done in 3D images is to swap the Z axis to the first index:
>>> a = a.transpose((2,0,1))
>>> a.shape
(20, 100, 100)
And now you can easily iterate over the Z axis:
>>> for slice in a:
do something
The slice here will be each of your 100x100 fractions of your 3D matrix. Additionally, by transpossing allows you to access each of the 2D slices directly by indexing the first axis. For example a[10] will give you the 11th 2D 100x100 slice.
Bonus: If you store the data contiguosly, without transposing (or converting to a contiguous array using a = np.ascontiguousarray(a.transpose((2,0,1))) the access to you 2D slices will be faster since they are mapped contiguosly in memory.
Obviously you want to get rid of the explicit for loop, but I think whether this is possible depends on what calculation you are doing with C. As a simple example,
a = np.zeros((100,100, 20))
a[:,:] = np.linspace(1,20,20) # example data: 1,2,3,.., 20 as "z" for every "x","y"
b = np.sum(a[:,:]**2, axis=2)
will fill the 100 by 100 array b with the sum of the squared "z" values of a, that is 1+4+9+...+400 = 2870.
If your inner calculation is sufficiently complex, and not amenable to vectorization, then your iteration structure is good, and does not contribute significantly to the calculation time
for (x,y), v in np.ndenumerate(b):
C = a[x,y,:]
...
for d in range(len(C)):
... # complex, not vectorizable calc
...
b[x,y] = some_value
There doesn't appear to be a special structure in the 1st 2 dimensions, so you could just as well think of it as 2D mapping on to 1D, e.g. mapping a (N,20) array onto a (N,) array. That doesn't speed up anything, but may help highlight the essential structure of the problem.
One step is to focus on speeding up that C to some_value calculation. There are functions like cumsum and cumprod that help you do sequential calculations on a vector. cython is also a good tool.
A different approach is to see if you can perform that internal calculation over the N values all at once. In other words, if you must iterate, it is better to do so over the smallest dimension.
In a sense this a non-answer. But without full knowledge of how you get some_value from C and d_n I don't think we can do more.
It looks like e can be calculated for all points at once:
e = 2.5 * float(math.pow(d_n[x,y] - d, 2)) + C[d] * 0.05
E = 2.5 * (d_n[...,None] - np.arange(a.shape[-1]))**2 + a * 0.05 # (100,100,20)
E.min(axis=-1) # smallest value along the last dimension
E.argmin(axis=-1) # index of where that min occurs
On first glance it looks like this E.argmin is the b value that you want (tweaked for some boundary conditions if needed).
I don't have realistic a and d_n arrays, but with simple test ones, this E.argmin(-1) matches your b, with a 66x speedup.
How can I improve the performance of such kind of functions mapping a 3D array to a 2D one?
Many functions in Numpy are "reduction" functions*, for example sum, any, std, etc. If you supply an axis argument other than None to such a function it will reduce the dimension of the array over that axis. For your code you can use the argmin function, if you first calculate e in a vectorized way:
d = np.arange(a.shape[2])
e = 2.5 * (d_n[...,None] - d)**2 + a*0.05
b = np.argmin(e, axis=2)
The indexing with [...,None] is used to engage broadcasting. The values in e are floating point values, so it's a bit strange to compare to sys.maxint but there you go:
I, J = np.indices(b.shape)
b[e[I,J,b] >= sys.maxint] = a.shape[2] - 1
* Strickly speaking a reduction function is of the form reduce(operator, sequence) so technically not std and argmin

numpy broadcast from first dimension

In NumPy, is there an easy way to broadcast two arrays of dimensions e.g. (x,y) and (x,y,z)? NumPy broadcasting typically matches dimensions from the last dimension, so usual broadcasting will not work (it would require the first array to have dimension (y,z)).
Background: I'm working with images, some of which are RGB (shape (h,w,3)) and some of which are grayscale (shape (h,w)). I generate alpha masks of shape (h,w), and I want to apply the mask to the image via mask * im. This doesn't work because of the above-mentioned problem, so I end up having to do e.g.
mask = mask.reshape(mask.shape + (1,) * (len(im.shape) - len(mask.shape)))
which is ugly. Other parts of the code do operations with vectors and matrices, which also run into the same issue: it fails trying to execute m + v where m has shape (x,y) and v has shape (x,). It's possible to use e.g. atleast_3d, but then I have to remember how many dimensions I actually wanted.
how about use transpose:
(a.T + c.T).T
numpy functions often have blocks of code that check dimensions, reshape arrays into compatible shapes, all before getting down to the core business of adding or multiplying. They may reshape the output to match the inputs. So there is nothing wrong with rolling your own that do similar manipulations.
Don't offhand dismiss the idea of rotating the variable 3 dimension to the start of the dimensions. Doing so takes advantage of the fact that numpy automatically adds dimensions at the start.
For element by element multiplication, einsum is quite powerful.
np.einsum('ij...,ij...->ij...',im,mask)
will handle cases where im and mask are any mix of 2 or 3 dimensions (assuming the 1st 2 are always compatible. Unfortunately this does not generalize to addition or other operations.
A while back I simulated einsum with a pure Python version. For that I used np.lib.stride_tricks.as_strided and np.nditer. Look into those functions if you want more power in mixing and matching dimensions.
as another angle: if you encounter this pattern frequently, it may be useful to create a utility function to enforce right-broadcasting:
def right_broadcasting(arr, target):
return arr.reshape(arr.shape + (1,) * (target.ndim - arr.ndim))
Although if there are only two types of input (already having 3 dims or having only 2), id say the single if statement is preferable.
Indexing with np.newaxis creates a new axis in that place. Ie
xyz = #some 3d array
xy = #some 2d array
xyz_sum = xyz + xy[:,:,np.newaxis]
or
xyz_sum = xyz + xy[:,:,None]
Indexing in this way creates an axis with shape 1 and stride 0 in this location.
Why not just decorate-process-undecorate:
def flipflop(func):
def wrapper(a, mask):
if len(a.shape) == 3:
mask = mask[..., None]
b = func(a, mask)
return np.squeeze(b)
return wrapper
#flipflop
def f(x, mask):
return x * mask
Then
>>> N = 12
>>> gs = np.random.random((N, N))
>>> rgb = np.random.random((N, N, 3))
>>>
>>> mask = np.ones((N, N))
>>>
>>> f(gs, mask).shape
(12, 12)
>>> f(rgb, mask).shape
(12, 12, 3)
Easy, you just add a singleton dimension at the end of the smaller array. For example, if xyz_array has shape (x,y,z) and xy_array has shape (x,y), you can do
xyz_array + np.expand_dims(xy_array, xy_array.ndim)

Categories

Resources