Jeff - do you know if it's possible to upload the Python helper files to a kernel on Kaggle? I'd prefer to use their hardware to train, but I want to use your utilities. Does this mean I have to train locally? Thanks again for this amazing tutorial!
This has not (yet) been a priority for me. But is on my list to test our and better support. At this point I have my utilities designed so that I can launch docker containers to AWS Fargate and execute several different trainings in parallel. Also not have to re-engineer the features each time (since my solutions are usually heavy on feature engineering. However, I do want to ultimately get this to something that can run as a kernal and be pip-installed.
Thanks for the video,But video is not much visible can you upload 1080p version of it.
Yes it could be clearer, but I think it can still be read. Sorry about that, I usually record at 1080, will do that next time!
This is a great example of clean and purposeful coding! Thank you, Jeff!
Thanks! this series is awesome... keep up the good work!
You are welcome! Next one is coming this Monday.
great stuff! hope they are still useful in years to come!
I see a lot of similarity with Kedro in your workflow, thanks for the advices, i"m gona mount this structure at Kedro and give a try!
Thank you Jeff, keep up the great work
Thanks you Jeff, thanks for sharing
Thanks Jeff for your valuable info
thank you jeff for your awesome videos ,should i do that github proccess on my laptop or i can do it on kaggle too
Could I please request to record your videos at least in HD quality? Thanks
So are you not using Jupyter Notebooks at all?
thanks so much, u r awesome
Jeff - do you know if it's possible to upload the Python helper files to a kernel on Kaggle? I'd prefer to use their hardware to train, but I want to use your utilities. Does this mean I have to train locally? Thanks again for this amazing tutorial!
This has not (yet) been a priority for me. But is on my list to test our and better support. At this point I have my utilities designed so that I can launch docker containers to AWS Fargate and execute several different trainings in parallel. Also not have to re-engineer the features each time (since my solutions are usually heavy on feature engineering. However, I do want to ultimately get this to something that can run as a kernal and be pip-installed.
@@HeatonResearch Thank you
Where do you take other people's code from? Kernels?
Kernels, blogs, and discussion forums on Kaggle itself.