Different Residual Size of PDE!!!

[ppandit95]

Dear Dr. Lu

Although I realize that this is a pretty silly question but I am solving a inverse problem wherein the coefficients of my PDE are functions of output for which I have written the PDE as -
def pde(x, M):
dM1_x = dde.grad.jacobian(M, x, i=0, j=0
dM2_x = dde.grad.jacobian(M, x, i=1, j=0)
M1 = M[:,0]
M2 = M[:,1]
D11 = tf.exp(-A1-A2*M1-A3*M2-A4*M1*M2-A5*M1*M1-A6*M2*M2-A7*M1*M1*M2-A8*M1*M2*M2)
D12 = tf.exp(-B1-B2*M1-B3*M2-B4*M1*M2-B5*M1*M1-B6*M2*M2-B7*M1*M1*M2-B8*M1*M2*M2)
D21 = tf.exp(-C1-C2*M1-C3*M2-C4*M1*M2-C5*M1*M1-C6*M2*M2-C7*M1*M1*M2-C8*M1*M2*M2)
D22 = tf.exp(-E1-E2*M1-E3*M2-E4*M1*M2-E5*M1*M1-E6*M2*M2-E7*M1*M1*M2-E8*M1*M2*M2)
R1_1 = JNi(x) + D11*dM1_x + D12*dM2_x
R1_2 = JCo(x) + D21*dM1_x + D22*dM2_x
return R1_1 , R1_2

Here,the issue I am facing is that in sample example inverse problem provided on github upon invoking the statement

f = model.predict(observe_x, operator=pde) print(np.array(f).shape)

I am getting a output size as (2,1493,1) where I have 2 residual components at each training data point while 1493 are the number of training data points while in my case from the PDE function that is elaborated above I am getting the shape as (2,1493,1493) . So, the query I have is that is there something I am doing incorrectly because I suspected that I should multiply vectors using tf.multiply but it doesn't work when implemented.So any help in this regard will be helpful.

Many Thanks
Pushkar

[ppandit95]

@lululxvi I did manage to resolve the above issue by simply changing the code
M1=M[:,0] ==> M1=M[:,0:1]
So can this thread be closed