CODEWITHSUNDEEP
pythontensorflowmachine-learninggoogle-colaboratory
rzaratx
rzaratx

Reputation: 824

Can Google Colab use local resources?

I am trying out Google Colab and wanted to know if I am able to use my local CPU, RAM, SSDs, and GPUs? I have tried to search a directory on my SSD but comes up empty. I found I have to upload my directories to my Google Drive and run the code:

from google.colab import drive
drive.mount('/content/drive')

Now that I have my directory established and I run my UNet I get the warning WARNING:tensorflow:There are non-GPU devices in `tf.distribute.Strategy`, not using nccl allreduce. and INFO:tensorflow:Using MirroredStrategy with devices ('/job:localhost/replica:0/task:0/device:CPU:0',) I am using Linux RedHat 8 and have install the Nvidia NCCL drivers. I do not get this warning when using Jupyter Notebook.

Before my first epoch begins my session crashes and gives the message Your session crashed after using all available RAM. Get more RAM. In the upper right hand corner Colab says RAM: 12.72 GB and Disk: 107.77. My system has a Terabyte SSD for storage and 128 GB of RAM. Is there a way to run Colab with my resources or do I have to use theirs?

Upvotes: 3

Views: 15719

Answers (1)

korakot
korakot

Reputation: 40828

You can use your own local resource by choosing "Local Runtime"

See documentation here:

https://research.google.com/colaboratory/local-runtimes.html

Upvotes: 4

Related Questions