General Questions on Algorithms

Do all libraries use the same way to undergo a certain algorithm?
For example:
In Pytorch Linear Regression calculates weights and bias for a polynomial expression. It optimises the weights and biases as per our target. Now,
If I use ScikitLearn, where it forms a fitline, are the behind the scene calculations the same?

Yes, the calculations happening behind the scenes are pretty much the same, although some specific details like initialization of weights, default optimizers & learning rates etc. may differ from framework to framework. Also, as far as I know fitline from scikit-learn does not work with batches of data, so you may not be able to use it for really large datasets.

It’s good to be familiar with both frameworks so that you can use the right framework for the problem at hand.

2 Likes

Thanks this was really helpful.

All pre-trained models are for 3channel images. How do I use this for gray scale images???
I found a link but that wasn’t really helpful. Or maybe very tough for me to understand.
Is there a way to use resnet34 or resnet50 or any other Pre-trained model for gray scale images.