I am debugging a neural network which has a torch.nn.functional.grid.sample operator inside. Using the Pycharm IDE, I can watch the values during debugging. My grid is a 1*15*2 tensor, here are the values in the first batch.
My input is a 1*128*16*16 tensor, here are the values in the first channel of the first batch:.
My output is 1*128*1*15 tensor, here are the values in the first channel of the first batch.
align_corners = False, mode = 'bilinear', padding_mode = 'zero'.
For gird coordinates (-1,-1), I can understand that the value(-4.74179) is sampled from 4 values on the top-left corner with 3 of them being the padded '0's and 1 of them being the value '-18.96716'.(-18.96716/4 = -4.74179).
But for other grid coordinates, I am confused. Taking the value '84.65594' for example, it's corresponding grid coordinate is (-0.45302, 0.53659). I firstly convert them from (-1,1) to (0,15) by adding 1 and then dividing by 2 and then multiplying 15(see official implementation). The converted coordinate is then (4.10235, 11.524425), Upon which I see the four values that should be sampled from are :
(x)44.20010---0.10235---------(y)26.68777
| | |
| | |
0.524425---(a,b)--------------------
| | |
| | |
(w)102.18765---------------------(z)30.03996
here are my calculation by hand step, Let:
a = 0.10235
b = 0.524425
x = 44.20010
y = 26.68777
z = 30.03996
w = 102.18765
The interpolated value should then be:
output = a*b*z + (1 - a)*(1 - b)*x + (1 - a)*b*w + (1-b)*a*y
= 0.10235*0.524425*30.03996 + (1-0.10235)*(1-0.524425)*44.20010 + (1-
0.10235)*0.524425*102.18765 + (1-0.524425)*0.10235*26.68777
= 69.8852865171
which isn't 84.65594, I cant't figure out how the value '84.65594' in the output is calculated, please help!
I answer my own question, it turns out that the inconsistency is due to the 'align_corners' flag. My way of calculation is actually under the case when 'align_corners' is true while in the program, this flag is set to be false. For how to calculate sample coordinates, please see this
Related
I have a custom (discrete) probability distribution defined somewhat in the form: f(x)/(sum(f(x')) for x' in a given discrete set X). Also, 0<=x<=1.
So I have been trying to implement it in python 3.8.2, and the problem is that the numerator and denominator both come out to be really small and python's floating point representation just takes them as 0.0.
After calculating these probabilities, I need to sample a random element from an array, whose each index may be selected with the corresponding probability in the distribution. So if my distribution is [p1,p2,p3,p4], and my array is [a1,a2,a3,a4], then probability of selecting a2 is p2 and so on.
So how can I implement this in an elegant and efficient way?
Is there any way I could use the np.random.beta() in this case? Since the difference between the beta distribution and my actual distribution is only that the normalization constant differs and the domain is restricted to a few points.
Note: The Probability Mass function defined above is actually in the form given by the Bayes theorem and f(x)=x^s*(1-x)^f, where s and f are fixed numbers for a given iteration. So the exact problem is that, when s or f become really large, this thing goes to 0.
You could well compute things by working with logs. The point is that while both the numerator and denominator might underflow to 0, their logs won't unless your numbers are really astonishingly small.
You say
f(x) = x^s*(1-x)^t
so
logf (x) = s*log(x) + t*log(1-x)
and you want to compute, say
p = f(x) / Sum{ y in X | f(y)}
so
p = exp( logf(x) - log sum { y in X | f(y)}
= exp( logf(x) - log sum { y in X | exp( logf( y))}
The only difficulty is in computing the second term, but this is a common problem, for example here
On the other hand computing logsumexp is easy enough to to by hand.
We want
S = log( sum{ i | exp(l[i])})
if L is the maximum of the l[i] then
S = log( exp(L)*sum{ i | exp(l[i]-L)})
= L + log( sum{ i | exp( l[i]-L)})
The last sum can be computed as written, because each term is now between 0 and 1 so there is no danger of overflow, and one of the terms (the one for which l[i]==L) is 1, and so if other terms underflow, that is harmless.
This may however lose a little accuracy. A refinement would be to recognize the set A of indices where
l[i]>=L-eps (eps a user set parameter, eg 1)
And then compute
N = Sum{ i in A | exp(l[i]-L)}
B = log1p( Sum{ i not in A | exp(l[i]-L)}/N)
S = L + log( N) + B
I have an image like that:
I have both the mask and the original image. I would like to calculate the colour temperature of ONLY the ducks region.
Right now, I'm iterating through each row and column of the image below and getting pixels where their values are not zero. But I think this isn't the right way to do this. Any suggestions?
What I did was:
xyzImg = cv2.cvtColor(resImage, cv2.COLOR_BGR2XYZ)
x,y,z = cv2.split(xyzImg)
xList=[]
yList=[]
zList=[]
rows=x.shape[0]
cols=x.shape[1]
for i in range(rows):
for j in range(cols):
if (x[i][j]!=0) and (y[i][j]!=0) and (z[i][j]!=0):
xList.append(x[i][j])
yList.append(y[i][j])
zList.append(z[i][j])
xAvg = np.mean(xList)
yAvg = np.mean(yList)
zAvg = np.mean(zList)
xs = xAvg / (xAvg + yAvg + zAvg)
ys = yAvg / (xAvg + yAvg + zAvg)
xyChrome = np.array([xs,ys])
But this is very slow and I don't think its right...
The simplest way would be to use cv2.mean() function.
It takes two arguments src (having 1 to 4 channels) and mask and returns a vector with mean values for individual channels.
Refer to cv2::mask
I must solve the Euler Bernoulli differential beam equation which is:
w’’’’(x) = q(x)
and boundary conditions:
w(0) = w(l) = 0
and
w′′(0) = w′′(l) = 0
The beam is as shown on the picture below:
beam
The continious force q is 2N/mm.
I have to use shooting method and scipy.integrate.odeint() func.
I can't even manage to start as i do not understand how to write the differential equation as a system of equation
Can someone who understands solving of differential equations with boundary conditions in python please help!
Thanks :)
The shooting method
To solve the fourth order ODE BVP with scipy.integrate.odeint() using the shooting method you need to:
1.) Separate the 4th order ODE into 4 first order ODEs by substituting:
u = w
u1 = u' = w' # 1
u2 = u1' = w'' # 2
u3 = u2' = w''' # 3
u4 = u3' = w'''' = q # 4
2.) Create a function to carry out the derivation logic and connect that function to the integrate.odeint() like this:
function calc(u, x , q)
{
return [u[1], u[2], u[3] , q]
}
w = integrate.odeint(calc, [w(0), guess, w''(0), guess], xList, args=(q,))
Explanation:
We are sending the boundary value conditions to odeint() for x=0 ([w(0), w'(0) ,w''(0), w'''(0)]) which calls the function calc which returns the derivatives to be added to the current state of w. Note that we are guessing the initial boundary conditions for w'(0) and w'''(0) while entering the known w(0)=0 and w''(0)=0.
Addition of derivatives to the current state of w occurs like this:
# the current w(x) value is the previous value plus the current change of w in dx.
w(x) = w(x-dx) + dw/dx
# others are calculated the same
dw(x)/dx = dw(x-dx)/dx + d^2w(x)/dx^2
# etc.
This is why we are returning values [u[1], u[2], u[3] , q] instead of [u[0], u[1], u[2] , u[3]] from the calc function, because u[1] is the first derivative so we add it to w, etc.
3.) Now we are able to set up our shooting method. We will be sending different initial boundary values for w'(0) and w'''(0) to odeint() and then check the end result of the returned w(x) profile to determine how close w(L) and w''(L) got to 0 (the known boundary conditions).
The program for the shooting method:
# a function to return the derivatives of w
def returnDerivatives(u, x, q):
return [u[1], u[2], u[3], q]
# a shooting funtion which takes in two variables and returns a w(x) profile for x=[0,L]
def shoot(u2, u4):
# the number of x points to calculate integration -> determines the size of dx
# bigger number means more x's -> better precision -> longer execution time
xSteps = 1001
# length of the beam
L= 1.0 # 1m
xSpace = np.linspace(0, L, xSteps)
q = 0.02 # constant [N/m]
# integrate and return the profile of w(x) and it's derivatives, from x=0 to x=L
return odeint(returnDerivatives, [ 0, u2, 0, u4] , xSpace, args=(q,))
# the tolerance for our results.
tolerance = 0.01
# how many numbers to consider for u2 and u4 (the guess boundary conditions)
u2_u4_maxNumbers = 1327 # bigger number, better precision, slower program
# you can also divide into separate variables like u2_maxNum and u4_maxNum
# these are already tested numbers (the best results are somewhere in here)
u2Numbers = np.linspace(-0.1, 0.1, u2_u4_maxNumbers)
# the same as above
u4Numbers = np.linspace(-0.5, 0.5, u2_u4_maxNumbers)
# result list for extracted values of each w(x) profile => [u2Best, u4Best, w(L), w''(L)]
# which will help us determine if the w(x) profile is inside tolerance
resultList = []
# result list for each U (or w(x) profile) => [w(x), w'(x), w''(x), w'''(x)]
resultW = []
# start generating numbers for u2 and u4 and send them to odeint()
for u2 in u2Numbers:
for u4 in u4Numbers:
U = []
U = shoot(u2,u4)
# get only the last row of the profile to determine if it passes tolerance check
result = U[len(U)-1]
# only check w(L) == 0 and w''(L) == 0, as those are the known boundary cond.
if (abs(result[0]) < tolerance) and (abs(result[2]) < tolerance):
# if the result passed the tolerance check, extract some values from the
# last row of the w(x) profile which we will need later for comaprisons
resultList.append([u2, u4, result[0], result[2]])
# add the w(x) profile to the list of profiles that passed the tolerance
# Note: the order of resultList is the same as the order of resultW
resultW.append(U)
# go through the resultList (list of extracted values from last row of each w(x) profile)
for i in range(len(resultList)):
x = resultList[i]
# both boundary conditions are 0 for both w(L) and w''(L) so we will simply add
# the two absolute values to determine how much the sum differs from 0
y = abs(x[2]) + abs(x[3])
# if we've just started set the least difference to the current
if i == 0:
minNum = y # remember the smallest difference to 0
index = 0 # remember index of best profile
elif y < minNum:
# current sum of absolute values is smaller
minNum = y
index = i
# print out the integral for w(x) over the beam
sum = 0
for i in resultW[index]:
sum = sum + i[0]
print("The integral of w(x) over the beam is:")
print(sum/1001) # sum/xSteps
This outputs:
The integral of w(x) over the beam is:
0.000135085272117
To print out the best profile for w(x) that we found:
print(resultW[index])
which outputs something like:
# w(x) w'(x) w''(x) w'''(x)
[[ 0.00000000e+00 7.54147813e-04 0.00000000e+00 -9.80392157e-03]
[ 7.54144825e-07 7.54142917e-04 -9.79392157e-06 -9.78392157e-03]
[ 1.50828005e-06 7.54128237e-04 -1.95678431e-05 -9.76392157e-03]
...,
[ -4.48774290e-05 -8.14851572e-04 1.75726275e-04 1.01560784e-02]
[ -4.56921910e-05 -8.14670764e-04 1.85892353e-04 1.01760784e-02]
[ -4.65067671e-05 -8.14479780e-04 1.96078431e-04 1.01960784e-02]]
To double check the results from above we will also solve the ODE using the numerical method.
The numerical method
To solve the problem using the numerical method we first need to solve the differential equations. We will get four constants which we need to find with the help of the boundary conditions. The boundary conditions will be used to form a system of equations to help find the necessary constants.
For example:
w’’’’(x) = q(x);
means that we have this:
d^4(w(x))/dx^4 = q(x)
Since q(x) is constant after integrating we have:
d^3(w(x))/dx^3 = q(x)*x + C
After integrating again:
d^2(w(x))/dx^2 = q(x)*0.5*x^2 + C*x + D
After another integration:
dw(x)/dx = q(x)/6*x^3 + C*0.5*x^2 + D*x + E
And finally the last integration yields:
w(x) = q(x)/24*x^4 + C/6*x^3 + D*0.5*x^2 + E*x + F
Then we take a look at the boundary conditions (now we have expressions from above for w''(x) and w(x)) with which we make a system of equations to solve the constants.
w''(0) => 0 = q(x)*0.5*0^2 + C*0 + D
w''(L) => 0 = q(x)*0.5*L^2 + C*L + D
This gives us the constants:
D = 0 # from the first equation
C = - 0.01 * L # from the second (after inserting D=0)
After repeating the same for w(0)=0 and w(L)=0 we obtain:
F = 0 # from first
E = 0.01/12.0 * L^3 # from second
Now, after we have solved the equation and found all of the integration constants we can make the program for the numerical method.
The program for the numerical method
We will make a FOR loop to go through the entire beam for every dx at a time and sum up (integrate) w(x).
L = 1.0 # in meters
step = 1001.0 # how many steps to take (dx)
q = 0.02 # constant [N/m]
integralOfW = 0.0; # instead of w(0) enter the boundary condition value for w(0)
result = []
for i in range(int(L*step)):
x= i/step
w = (q/24.0*pow(x,4) - 0.02/12.0*pow(x,3) + 0.01/12*pow(L,3)*x)/step # current w fragment
# add up fragments of w for integral calculation
integralOfW += w
# add current value of w(x) to result list for plotting
result.append(w*step);
print("The integral of w(x) over the beam is:")
print(integralOfW)
which outputs:
The integral of w(x) over the beam is:
0.00016666652805511192
Now to compare the two methods
Result comparison between the shooting method and the numerical method
The integral of w(x) over the beam:
Shooting method -> 0.000135085272117
Numerical method -> 0.00016666652805511192
That's a pretty good match, now lets see check the plots:
From the plots it's even more obvious that we have a good match and that the results of the shooting method are correct.
To get even better results for the shooting method increase xSteps and u2_u4_maxNumbers to bigger numbers and you can also narrow down the u2Numbers and u4Numbers to the same set size but a smaller interval (around the best results from previous program runs). Keep in mind that setting xSteps and u2_u4_maxNumbers too high will cause your program to run for a very long time.
You need to transform the ODE into a first order system, setting u0=w one possible and usually used system is
u0'=u1,
u1'=u2,
u2'=u3,
u3'=q(x)
This can be implemented as
def ODEfunc(u,x): return [ u[1], u[2], u[3], q(x) ]
Then make a function that shoots with experimental initial conditions and returns the components of the second boundary condition
def shoot(u01, u03): return odeint(ODEfunc, [0, u01, 0, u03], [0, l])[-1,[0,2]]
Now you have a function of two variables with two components and you need to solve this 2x2 system with the usual methods. As the system is linear, the shooting function is linear as well and you only need to find the coefficients and solve the resulting linear system.
How can I transform the blue curve values into linear (red curve)? I am doing some tests in excel, but basically I have those blue line values inside a 3D App that I want to manipulate with python so I can make those values linear. Is there any mathematical approach that I am missing?
The x axis goes from 0 to 90, and the y axis from 0 to 1.
For example: in the middle of the graph the blue line gives me a value of "0,70711", and I know that in linear it is "0,5". I was wondering if there's an easy formula to transform all the incoming non-linear values into linear.
I have no idea what "formula" is creating that non-linear blue line, also ignore the yellow line since I was just trying to "reverse engineer" to see if would lead me to any conclusion.
Thank you
Find a linear function y = ax + b that for x = 0 gives the value 1 and for x = 90 gives 0, just like the function that is represented by a blue curve.
In that case, your system of equations is the following:
1 = b // for x = 0
0 = a*90 + b // for x = 90
Solution provided by solver is the following : { a = -1/90, b = 1 }, the red linear function will have form y = ax + b, we put the values of a and b we found from the solver and we discover that the linear function you are looking for is y = -x/90 + 1 .
The tool I used to solve the system of equations:
http://wims.unice.fr/wims/en_tool~linear~linsolver.en.html
What exactly do you mean? You can calculate points on the red line like this:
f(x) = 1-x/90
and the point then is (x,f(x)) = (x, 1-x/90). But to be honest, I think your question is still rather unclear.
So I was looking at this question:
Matlab - Standard Deviation of Cartesian Points
Which basically answers my question, except the problem is I have xyz, not xy. So I don't think Ax=b would work in this case.
I have, say, 10 Cartesian points, and I want to be able to find the standard deviation of these points. Now, I don't want standard deviation of each X, Y and Z (as a result of 3 sets) but I just want to get one number.
This can be done using MATLAB or excel.
To better understand what I'm doing, I have this desired point (1,2,3) and I recorded (1.1,2.1,2.9), (1.2,1.9,3.1) and so on. I wanted to be able to find the variability of all the recorded points.
I'm open for any other suggestions.
If you do the same thing as in the other answer you linked, it should work.
x_vals = xyz(:,1);
y_vals = xyz(:,2);
z_vals = xyz(:,3);
then make A with 3 columns,
A = [x_vals y_vals ones(size(x_vals))];
and
b = z_vals;
Then
sol=A\b;
m = sol(1);
n = sol(2);
c = sol(3);
and then
errs = (m*x_vals + n*y_vals + c) - z_vals;
After that you can use errs just as in the linked question.
Randomly clustered data
If your data is not expected to be near a line or a plane, just compute the distance of each point to the centroid:
xyz_bar = mean(xyz);
M = bsxfun(#minus,xyz,xyz_bar);
d = sqrt(sum(M.^2,2)); % distances to centroid
Then you can compute variability anyway you like. For example, standard deviation and RMS error:
std(d)
sqrt(mean(d.^2))
Data about a 3D line
If the data points are expected to be roughly along the path of a line, with some deviation from it, you might look at the distance to a best fit line. First, fit a 3D line to your points. One way is using the following parametric form of a 3D line:
x = a*t + x0
y = b*t + y0
z = c*t + z0
Generate some test data, with noise:
abc = [2 3 1]; xyz0 = [6 12 3];
t = 0:0.1:10;
xyz = bsxfun(#plus,bsxfun(#times,abc,t.'),xyz0) + 0.5*randn(numel(t),3)
plot3(xyz(:,1),xyz(:,2),xyz(:,3),'*') % to visualize
Estimate the 3D line parameters:
xyz_bar = mean(xyz) % centroid is on the line
M = bsxfun(#minus,xyz,xyz_bar); % remove mean
[~,S,V] = svd(M,0)
abc_est = V(:,1).'
abc/norm(abc) % compare actual slope coefficients
Distance from points to a 3D line:
pointCentroidSeg = bsxfun(#minus,xyz_bar,xyz);
pointCross = cross(pointCentroidSeg, repmat(abc_est,size(xyz,1),1));
errs = sqrt(sum(pointCross.^2,2))
Now you have the distance from each point to the fit line ("error" of each point). You can compute the mean, RMS, standard deviation, etc.:
>> std(errs)
ans =
0.3232
>> sqrt(mean(errs.^2))
ans =
0.7017
Data about a 3D plane
See David's answer.