Can I get the same results for the same dataset

[JackieShui]

When I use this toolkit for the same dataset over and over agian and I got different models. How can I get the same results for the same dataset?

[sharpe5]

If the random seed is the same, then results will be the same. You may have to set the random seed in more than one place.

On Thu, 13 Sep 2018 09:04 JackieShui, [email protected] wrote:

When I use this toolkit for the same dataset over and over agian and I got different models. How can I get the same results for the same dataset?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/ClimbsRocks/auto_ml/issues/415, or mute the thread https://github.com/notifications/unsubscribe-auth/ABOypPoM24wtBmuc5V3TaN7ZR-UUMmMHks5uahFMgaJpZM4Wmw22 .

[JackieShui]

If the random seed is the same, then results will be the same. You may have to set the random seed in more than one place. … On Thu, 13 Sep 2018 09:04 JackieShui, @.***> wrote: When I use this toolkit for the same dataset over and over agian and I got different models. How can I get the same results for the same dataset? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#415>, or mute the thread https://github.com/notifications/unsubscribe-auth/ABOypPoM24wtBmuc5V3TaN7ZR-UUMmMHks5uahFMgaJpZM4Wmw22 .

Does the auto_ml provide the params to set the random seed?

[sharpe5]

I imagine is does, try huntisng around. Python also has some settings for random seeds for some of the base libraries, there are two major ones, from memory. Just use trial and error until you get two identical runs.

On Thu, 13 Sep 2018 09:32 JackieShui, [email protected] wrote:

If the random seed is the same, then results will be the same. You may have to set the random seed in more than one place. … <#m_1876121903311457515_> On Thu, 13 Sep 2018 09:04 JackieShui, @.***> wrote: When I use this toolkit for the same dataset over and over agian and I got different models. How can I get the same results for the same dataset? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#415>, or mute the thread https://github.com/notifications/unsubscribe-auth/ABOypPoM24wtBmuc5V3TaN7ZR-UUMmMHks5uahFMgaJpZM4Wmw22 .

Does the auto_ml provide the params to set the random seed?

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/ClimbsRocks/auto_ml/issues/415#issuecomment-420927658, or mute the thread https://github.com/notifications/unsubscribe-auth/ABOypJT7F26YlRpwqkRCYuqDm7i8KLVLks5uahgFgaJpZM4Wmw22 .

[JackieShui]

I imagine is does, try huntisng around. Python also has some settings for random seeds for some of the base libraries, there are two major ones, from memory. Just use trial and error until you get two identical runs. …

Thank you! I'll try it.