search

Speedml / speedml

Speedml is a Python package to speed start machine learning projects.
MIT License
211 stars 32 forks source link

SpeedML codes not Executing #25

Closed motisingh65 closed 7 years ago

motisingh65 commented 7 years ago

I have installed gcc, xgboost, speedml all installed on my machine - however, am still not able to run the speedml codes on my notebooks....is it that I can run SpeedML on projects that are hosted on GitHub only

manavsehgal commented 7 years ago

@motisingh65 can you please share your notebook on GitHub so we can have a look. Or, share the code snippet and error message here.

motisingh65 commented 7 years ago

Manav,

Appreciate your response

I have resolved this one as well. I re-downloaded the speedml-0.9.3 package and it worked on Titanic dataset, though I have yet to test it on other dataset. Will let you know, if I hit a roadblock.

Do you think there should be any issues while using Speedml syntax with any other external data sets?

Regards

Moti

From: Manav Sehgal notifications@github.com Sent: Sunday, June 18, 2017 10:42 AM To: Speedml/speedml Cc: motisingh65; Mention Subject: Re: [Speedml/speedml] SpeedML codes not Executing (#25)

@motisingh65https://github.com/motisingh65 can you please share your notebook on GitHub so we can have a look. Or, share the code snippet and error message here.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/Speedml/speedml/issues/25#issuecomment-309257236, or mute the threadhttps://github.com/notifications/unsubscribe-auth/Aav0Ddei45TmWTmuO3XIZXdz5oefDAstks5sFLGpgaJpZM4N9OcB.

manavsehgal commented 7 years ago

Hi @motisingh65, The Speedml API has been tested on few competitions from Kaggle. Mostly classification problems with a single target variable. I am opening up the API to easily expose underlying packages like pandas, xgboost, numpy, etc. so that you have same flexibility when using the API as you have when using these packages directly. This should mean wider applicability across datasets. However, you may still find edge cases which API may not handle. Please report these here and I will try and address these in a future release. Thanks for using the API and sharing your experience.