search

ravin-d-27 / PyDeepFlow

This is my own Deep Learning Package, optimized for performing Deep Learning Tasks and easy to learn and integrate into projects.
https://pypi.org/project/pydeepflow/
MIT License
18 stars 8 forks source link

Implement Cross-Validation Support #16

Closed ravin-d-27 closed 1 month ago

ravin-d-27 commented 1 month ago

Add support for k-fold cross-validation to ensure the model generalizes well across different subsets of data.

Objective:
Improve model evaluation and robustness against overfitting.

Implementation Suggestions:

GGital commented 1 month ago

Hello @ravin-d-27! I would like to contribute in your repo in this topic. I would like to know if you still need additional functions or details as I see you already implemented the CrossValidator class. You can also suggest me other features if wanted!

ravin-d-27 commented 1 month ago

Hi @GGital ! Thank you for your interest! I want to test if this functionality is working or not. So could you please create a file called cross_validation_runner.py and check whether it's working or not?

If it's working, make a PR of cross_validation_runner.py file, else fix the problem and make a PR.

GGital commented 1 month ago

Hi @ravin-d-27! Thank you for your reply! I would like to ask you few questions before I proceed this.

  1. I want to know if there is any edits in the requirements.txt file. I can't use those dependencies to run runner.py.
  2. There is already a K-fold usage in runner.py at the moment. I don't know if there is still any requirement for me to work with this. If there is no need for the cross_validation_runner.py anymore, you can assign me another task as I am finding a repo about DL/ML to contribute with in this Hacktoberfest.
  3. Can you also assign me for this issue as well?
ravin-d-27 commented 1 month ago

Hi @GGital. There are no edits as of now. Only thing is if you are using PyDeepFlow with the help of GPU, you need to install cupy.

pip install cupy.

Other than that the requirements.txt is correct and functional. To avoid conflicts with your other dependencies in your python, consider using a virtual environment.

I also understand that there is no need for this issue as most of the things are already implemented. Testing is the only part which is remaining (Which I will do it).

So, I would like to assign another issue to you. Please check out the other issues and let me know which one you would like to work.

Thank you for your interest !