How can I use h2o4gpu on Google Colab?

607 Views Asked by At

I need to train a model (Random Forest) using scikit learn and I wanted to run it on Google Colab using the GPU option. I know I can't directly use sklearn for this purpose so I installed h2o4gpu on Google Colab as a drop-in replacement, setting all the parameters of the model to use it with the GPU. Everything seems to work fine, but then Google Colab tell me I'm not actually using the GPU (it asks me to switch to a standard environment since I'm not using the GPU). Unfortunately,I'm not able to train the model in the standard environment since it requires too much resources and I get a memory error.

So, my question is, why isn't it working? How can I correctly set up h2o4gpu so that I can use the Google Colab GPU option?

1

There are 1 best solutions below

0
On

Hi my friend I check it and made gits about it.

Install h2o4gpu and tpot on google colab (GPU)

  1. export some env variable
  2. install linux packges
  3. uninstall sklearn and install python packges
  4. enjoy fast auto ML with gpu

I hope this can help you to run your code easier.