Tanay Mehta

Results 35 comments of Tanay Mehta

> @heytanay you should be able to use scikit-learn's cross validators directly with a cuML model. Can you try that and let me know if it works? Hi, If I...

@beckernick I don't have a specific dataset in mind but there have been a few instances on Kaggle as well as off kaggle where I was dealing with really large...

Hi @beckernick, @divyegala Is this something that can be worked upon then? I would love to do the implementation of cross-validators if I have the green light.

@divyegala Thanks! I will start working on it and open a draft PR post-haste. I wanted to clear out a doubt: I don't suppose I'll have to write CUDA kernels...

@divyegala I've opened a draft PR [here](https://github.com/rapidsai/cuml/pull/4685), currently, I have just copied all the necessary functions as-it-is from the scikit-learn source code to cuml, I will be adapting them to...

Hi @divyegala, I think I will be needing [`dok_matrix`](https://docs.scipy.org/doc/scipy/reference/generated/scipy.sparse.dok_matrix.html) implemented in cupyx to complete the implementation of KFolds. Would love to hear your suggestions on how to proceed further.

Hi @divyegala, really sorry for replying so late, I was busy with my Undergrad finals, will start working on it again now!