We Keep Coding

torch gather using two index arrays

The goal is to extract a random 2x5 patch from a 5x10 image, and do so randomly for all images in a batch. Looking to write a faster implementation that avoids for loops. Haven't been able to figure out how to use the torch .gather operation with two index arrays (idx_h and idx_w in code example).
Naive for loop:
import torch
b = 3 # batch size
h = 5 # height
w = 10 # width
crop_border = (3, 5) # number of pixels (height, width) to crop
x = torch.arange(b * h * w).reshape(b, h, w)
print(x)
dh_ = torch.randint(0, crop_border[0], size=(b,))
dw_ = torch.randint(0, crop_border[1], size=(b,))
_dh = h - (crop_border[0] - dh_)
_dw = w - (crop_border[1] - dw_)
idx_h = torch.stack([torch.arange(d_, _d) for d_, _d in zip(dh_, _dh)])
idx_w = torch.stack([torch.arange(d_, _d) for d_, _d in zip(dw_, _dw)])
print(idx_h, idx_w)
new_shape = (b, idx_h.shape[1], idx_w.shape[1])
cropped_x = torch.empty(new_shape)
for batch in range(b):
for height in range(idx_h.shape[1]):
for width in range(idx_w.shape[1]):
cropped_x[batch, height, width] = x[
batch, idx_h[batch, height], idx_w[batch, width]
]
print(cropped_x)

Index arrays needed to be repeated and reshaped to work with gather operation. Fast_crop code based pytorch discussion: https://discuss.pytorch.org/t/similar-to-torch-gather-over-two-dimensions/118827
def fast_crop(x, idx1, idx2):
"""
Compute
x: N x B x V
idx1: N x K matrix where idx1[i, j] is between [0, B)
idx2: N x K matrix where idx2[i, j] is between [0, V)
Return:
cropped: N x K matrix where y[i, j] = x[i, idx1[i,j], idx2[i,j]]
"""
x = x.contiguous()
assert idx1.shape == idx2.shape
lin_idx = idx2 + x.size(-1) * idx1
x = x.view(-1, x.size(1) * x.size(2))
lin_idx = lin_idx.view(-1, lin_idx.shape[1] * lin_idx.shape[2])
cropped = x.gather(-1, lin_idx)
return cropped.reshape(idx1.shape)
idx1 = torch.repeat_interleave(idx_h, idx_w.shape[1]).reshape(new_shape)
idx2 = torch.repeat_interleave(idx_w, idx_h.shape[1], dim=0).reshape(new_shape)
cropped = fast_crop(x, idx1, idx2)
(cropped == cropped_x).all()
Using realistic numbers for b = 100, h = 100, w = 130 and crop_border = (40, 95), a 10 trial run takes the for loop 32s while fast_crop only 0.043s.

Python - ValueError: operands could not be broadcast together with shapes (17,90) (17,)

I am trying to implement logistic regression with regularization in Python using optimize.minimize from the SciPy library. Here is my code:
import pandas as pd
import numpy as np
from scipy import optimize
l = 0.1 # lambda
def sigmoid(z):
return 1 / (1 + np.exp(-z))
def cost_function_logit(theta, X, y, l):
h = sigmoid(X # theta)
# cost
J = -1 / m * (y.T # np.log(h)
+ (1 - y).T # np.log(1 - h)) \
+ l / (2 * m) * sum(theta[1:] ** 2)
# gradient
a = 1 / m * X.T # (h - y)
b = l / m * theta
grad = a + b
grad[0] = 1 / m * sum(h - y)
return J, grad
data = pd.read_excel('Data.xlsx')
X = data.drop(columns = ['healthy'])
m, n = X.shape
X = X.to_numpy()
X = np.hstack([np.ones([m, 1]), X])
y = pd.DataFrame(data, columns = ['healthy'])
y = y.to_numpy()
initial_theta = np.zeros([n + 1, 1])
options = {'maxiter': 400}
res = optimize.minimize(cost_function_logit,
initial_theta,
(X, y, l),
jac = True,
method = 'TNC',
options = options)
An error occurs on the line where I use optimize.minimize. The last two lines of the error are as follows:
grad = a + b
ValueError: operands could not be broadcast together with shapes (17,90) (17,)
I have checked the type and dimensions of X, y and theta, and they seem correct to me.
>>> type(X)
<class 'numpy.ndarray'>
>>> type(y)
<class 'numpy.ndarray'>
>>> type(theta)
<class 'numpy.ndarray'>
>>> X.shape
(90, 17)
>>> y.shape
(90, 1)
>>> theta.shape
(17, 1)
The error says a is a (17,90) matrix but based on my calculations it should be a (17,1) vector. Does anyone know where I went wrong?

I found a solution. Apparently, optimize.minimize didn't like that y and theta had shapes (90,1) and (17,1), respectively. I converted their shape to (90,) and (17,) and the error message went away.
In terms of code, I changed
initial_theta = np.zeros([n + 1, 1])
to just this:
initial_theta = np.zeros([n + 1])
and I added the following line:
y = np.reshape(y, [m])
Thanks to those who tried to help me.

The elements of a are 90 dimensional vectors, whereas the elements of b are numbers. I'm not totally sure what you're trying to do, but if you want to add vectors, they need to have the same shape. If you want to add the thing in b to each element in a row-wise you can do
grad = a + np.stack((b,) * a.shape[1], axis=-1)
but I'm assuming you just are messing up constructing a.

How to multiply N vectors by N matrices using numpy?

I have a matrix M of shape (N, L) and a 3D tensor P of shape (N, L, K). I want to get matrix V of shape (N, K) where V[i] = M[i] # P[i]. I can do it with for loop but that's inefficient, I want to do it with a single or few operations so that it would run in parallel on CUDA.
I tried just multiplying it like so
V = M # P
but that results in a 3D tensor where V[i, j] = M[j] # P[i].
np.diagonal(M # P).T is basically what I want, but calculating it like that wastes a lot of computation.

You could use np.einsum:
>>> M = np.random.rand(5, 2)
>>> P = np.random.rand(5, 2, 3)
>>> V = np.einsum('nl,nlk->nk', M, P)
>>> V.shape
(5, 3)

How can I code a matrix within a matrix using a loop?

So I have this 3x3 G matrix (not shown here, it's irrelevant to my problem) that I created using the two variables u (a vector, x - y) and the scalar k. x_j = (x_1 (j), x_2 (j), x_3 (j)) and y_j = (y_1 (j), y_2 (j), y_3 (j)). alpha_j is a 3x3 matrix. The A matrix is block diagonal matrix of size 3nx3n. I am having trouble with the W matrix. How do I code a matrix of size 3nx3n, where the (i,j)th block is the 3x3 matrix given by alpha_i*G_[ij]*alpha_j?? I am lost.
My alpha_j matrix also seems to be having some trouble. The loop keeps throwing me the error, "only length-1 arrays can be converted to Python scalars." pls help :/
def W(x, y, k, alpha, A):
u = x - y
n = x.shape[0]
W = np.zeros((3*n, 3*n))
for i in range(0, n-1):
for j in range(0, n-1):
#u = -np.array([[x[i,0] - x[j,0]], [x[i,1] - x[j,1]], [0]]) ??
W[i][j] = (alpha_j(alpha, A) * G(u, k) * alpha_j(alpha, A))
W[i][i] = np.zeros((n, n))
return W
def alpha_j(a, A):
alph = np.array([[0,0,0],[0,0,0],[0,0,0]],complex)
rho = np.random.rand(3,1)
for i in range(0, 2):
for j in range(0, 2):
alph[i][j] = (rho[i] * a * A[i][j])
return alph
#-------------------------------------------------------------------
x1 = np.array([[1], [2], [0]])
y1 = np.array([[4], [5], [0]])
# SYSTEM PARAMETERS
# incoming Wave angle
theta = 0 # can range from [0, 2pi)
# susceptibility
chi = 10 + 1j
# wavelength
lam = 0.5 # microns (values between .4-.7)
# frequency
k = (2 * np.pi)/lam # 1/microns
# volume
V_0 = (0.05)**3 # microns^3
# incoming wave vector
K = k * np.array([[0], [np.sin(theta)], [np.cos(theta)]])
# polarization vector
vecinc = np.array([[1], [0], [0]]) # (can choose any vector perpendicular to K)
# for the fixed alpha case
alpha = (V_0 * 3 * chi)/(chi + 3)
# 3 x 3 matrix
A = np.matlib.identity(3) # could be any symmetric matrix,
#-------------------------------------------------------------------
# TEST FUNCTIONS
test = G((x1-y1), k)
print(test)
w = W(x1, y1, k, alpha, A)
print(w)
Sometimes my W loops throws me the error, "can't set an array element with a sequence." But I need to set each array element in this arbitrary matrix W to the 3x3 matrix created by multiplying alpha by G...

To your question of how to create a new array with a block for each element, the following should do the trick:
G = np.random.random([3,3])
result = np.zeros([9,9])
num_blocks = 3
a = np.random.random([3,3])
b = np.random.random([3,3])
for i in range(G.shape[0]):
for j in range(G.shape[1]):
block_result = a*G[i,j]*b
for k in range(num_blocks):
for l in range(num_blocks):
result[3*i + k, 3*j + l] = block_result[i, j]
You should be able to generalize from there. I hope I've understood correctly.
EDIT: It looks like I haven't understood correctly. I'm leaving it in hopes it spurs you to an answer. The general idea is to generate ranges of indices to operate on, and then just operate on them directly. Slicing might be helpful, too.
Ah, you asked how to create a diagonal filled with blocks. In that case:
num_diagonal_blocks = 3 # for example
for block_dim in range(num_diagonal_blocks)
# do your block calculation...
for k in range(G.shape[0]):
for l in range(G.shape[1]):
result[3*block_dim + k, 3*block_dim + l] = # assign to element of block
I think that's nearly it.

Generate matrices for pairs of values in Numpy

I have a 3D array (a 2D array of vectors), of which I want to transform each vector with a rotation matrix. The rotations are in two separate 2D arrays of radians angle values called cols and rows.
I've been able to have NumPy compute the angles for me already, without a Python loop. Now I'm looking for a way to have NumPy generate the rotation matrices, too, hopefully resulting in a great performance boost.
size = img.shape[:2]
# Create an array that assigns each pixel the percentage of
# the correction (value between -1 and 1, distributed linearly).
cols = np.array([np.arange(size[1]) for __ in range(size[0])]) / (size[1] - 1) * 2 - 1
rows = np.array([np.arange(size[0]) for __ in range(size[1])]).T / (size[0] - 1) * 2 - 1
# Atan distribution based on F-number and Sensor size.
cols = np.arctan(sh * cols / (2 * f))
rows = np.arctan(sv * rows / (2 * f))
### This is the loop that I would like to remove and find a
### clever way to make NumPy do the same operation natively.
for i in range(size[0]):
for j in range(size[1]):
ah = cols[i,j]
av = rows[i,j]
# Y-rotation.
mat = np.matrix([
[ np.cos(ah), 0, np.sin(ah)],
[0, 1, 0],
[-np.sin(ah), 0, np.cos(ah)]
])
# X-rotation.
mat *= np.matrix([
[1, 0, 0],
[0, np.cos(av), -np.sin(av)],
[0, np.sin(av), np.cos(av)]
])
img[i,j] = img[i,j] * mat
return img
Is there any clever way to rewrite the loop in NumPy operations?

(Let's assume the shape of img be (a, b, 3).)
Firstly, cols and rows does not need to be fully expanded to (a, b) (you could write cols[j] instead of cols[i,j]). And they can be easy generated using np.linspace:
cols = np.linspace(-1, 1, size[1]) # shape: (b,)
rows = np.linspace(-1, 1, size[0]) # shape: (a,)
cols = np.arctan(sh * cols / (2*f))
rows = np.arctan(sv * rows / (2*f))
Then we get precalculate the components of the matrices.
# shape: (b,)
cos_ah = np.cos(cols)
sin_ah = np.sin(cols)
zeros_ah = np.zeros_like(cols)
ones_ah = np.ones_like(cols)
# shape: (a,)
cos_av = np.cos(rows)
sin_av = np.sin(rows)
zeros_av = np.zeros_like(rows)
ones_av = np.ones_like(rows)
And then construct the rotation matrices:
# shape: (3, 3, b)
y_mat = np.array([
[cos_ah, zeros_ah, sin_ah],
[zeros_ah, ones_ah, zeros_ah],
[-sin_ah, zeros_ah, cos_ah],
])
# shape: (3, 3, a)
x_mat = np.array([
[ones_av, zeros_av, zeros_av],
[zeros_av, cos_av, -sin_av],
[zeros_av, sin_av, cos_av],
])
Now let's see. If we have a loop we would write:
for i in range(size[0]):
for j in range(size[1]):
img[i, j, :] = img[i, j, :] # y_mat[:, :, j] # x_mat[:, :, i]
or, if we expand out the matrix multiplications:
This can be handled nicely using np.einsum (note the i,j,k,m,n corresponds exactly like the equation above):
img = np.einsum('ijk,kmj,mni->ijn', img, y_mat, x_mat)
To summarize:
size = img.shape[:2]
cols = np.linspace(-1, 1, size[1]) # shape: (b,)
rows = np.linspace(-1, 1, size[0]) # shape: (a,)
cols = np.arctan(sh * cols / (2*f))
rows = np.arctan(sv * rows / (2*f))
cos_ah = np.cos(cols)
sin_ah = np.sin(cols)
zeros_ah = np.zeros_like(cols)
ones_ah = np.ones_like(cols)
cos_av = np.cos(rows)
sin_av = np.sin(rows)
zeros_av = np.zeros_like(rows)
ones_av = np.ones_like(rows)
y_mat = np.array([
[cos_ah, zeros_ah, sin_ah],
[zeros_ah, ones_ah, zeros_ah],
[-sin_ah, zeros_ah, cos_ah],
])
x_mat = np.array([
[ones_av, zeros_av, zeros_av],
[zeros_av, cos_av, -sin_av],
[zeros_av, sin_av, cos_av],
])
return np.einsum('ijk,kmj,mni->ijn', img, y_mat, x_mat)