Computing the Pearson Correlation coefficient between a set of model weights trained in Keras

Question

I am training an ImageNet-pretrained VGG-16 model in Keras with two different hyperparameter settings. I prefer to see if there exists a linear relationship between the two sets of model weights. For simplicity, I have taken two VGG-16 models just to check if my approach works. I have computed the layer weights for each model and stored them as a 1-dimensional array, each having a shape of (1, 512). I wish to measure the Pearson Correlation coefficient between these two model weights as used in this study https://arxiv.org/pdf/1910.08475v2.pdf. My code is as shown below:

import os
from tensorflow import keras
from tensorflow.keras import backend as K
from tensorflow.keras.layers import *
from tensorflow.keras import applications
from tensorflow.keras.models import Model
from scipy.stats import pearsonr
from numpy import array

model_input = (224,224,3)
model1 = applications.VGG16(include_top=False,
                           weights='imagenet',
                           input_shape=model_input)
model1.summary()

n_layers1 = len(model1.get_weights())
for layer1 in range(n_layers1):
    layer_weights1 = array([model1.get_weights()[layer1]]) # shape [1 x 512]
 

model2 = applications.VGG16(include_top=False,
                           weights='imagenet',
                           input_shape=model_input)
model2.summary()
n_layers2 = len(model2.get_weights())
for layer2 in range(n_layers2):
    layer_weights2 = array([model2.get_weights()[layer2]]) # shape [1 x 512]

# calculate Pearson's correlation
corr, _ = pearsonr(layer_weights1, layer_weights2)
print('Pearsons correlation: %.3f' % corr)

I am getting the following error:

 File "C:\Users\AppData\Local\Temp\2/ipykernel_16136/1139308701.py", line 1, in <module>
    corr, _ = pearsonr(layer_weights1, layer_weights2)

  File "C:\Users\AppData\Local\Continuum\anaconda3\envs\tf2.7\lib\site-packages\scipy\stats\stats.py", line 4016, in pearsonr
    raise ValueError('x and y must have length at least 2.')

ValueError: x and y must have length at least 2.

Is my approach correct?

shiva · Accepted Answer · 2023-01-23 19:58:49Z

I found a solution to my problem. It is the shape mismatch error due to the syntax for the computation of the Pearson correlation coefficient. The revised code is as shown below:

import os
from tensorflow import keras
from tensorflow.keras import backend as K
from tensorflow.keras.layers import *
from tensorflow.keras import applications
from tensorflow.keras.models import Model
from scipy.stats import pearsonr
from numpy import array
import numpy as np

model_input = (224,224,3)
model1 = applications.VGG16(include_top=False,
                           weights='imagenet',
                           input_shape=model_input)
model1.summary()
n_layers1 = len(model1.get_weights())
for layer1 in range(n_layers1):
    layer_weights1 = array([model1.get_weights()[layer1]])
 

model2 = applications.VGG16(include_top=False,
                           weights='imagenet',
                           input_shape=model_input)
model2.summary()
n_layers2 = len(model2.get_weights())
for layer2 in range(n_layers2):
    layer_weights2 = array([model2.get_weights()[layer2]])

# calculate Pearson's correlation
x1 = layer_weights1.transpose()
x2 = layer_weights1.transpose()
x11 = x1.reshape(-1)
x22 = x2.reshape(-1)
corr, sig = pearsonr(x11, x22)
print('Pearsons correlation: %.4f' % corr)

Collectives™ on Stack Overflow

Computing the Pearson Correlation coefficient between a set of model weights trained in Keras

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related