How to create dataset to run DL algorithm on our dataset
That is really beyond the scope of this course, but there is some information on the forums here that will get you started. You can start by studying how the existing datasets are packaged. E.g. look at the Course 1 Week 2 Logistic Regression assignment. If you study the logic that loads the dataset, you’ll find where the files are and that they are packaged as “h5” files. Look at the source code for the load_dataset routine, which is in the file lr_utils.py. Here’s a thread on the forum in which one of our fellow students did the research and wrote some sample code for creating an h5 file containing images like the ones we used in C1 W2 A2.
Of course there is no reason you have to choose h5 as your file format. A lot of projects use “zip” files or “tar” files. Just packaging up a set of images as a zip file is easy. There are python libraries for dealing with just about any file format. Just google “python create zip file” or whatever your preferred format is. But the one thing you will need to do is figure out how you’re going to encode the labels for your images.
Thank you sir for your immediate reply.
I did try out fetching the h5 format…but got stuck up.
Will try once again and revert back.
Thank you very much once again.
Hello Sir, thanks for the thread as to create h5 dataset. But the dataset created is for binary dataset, I need dataset for multiclass. I tried to extend the same for multiclass…but it not working. Kindly help me out in this regard.
Hi, Sunitha.
You probably already noticed, but I responded on the other thread where you posted this question.