Open diliptmonson opened 4 years ago
In the fast ai setup guide for GCP, https://course.fast.ai/start_gcp.html
it was recommended to use us-west1-b. However Nvidia Tesla P4 is not available in this instance and is available in us-central1-a.
Further, in instance creation code, Nvidia tesla p100 was provided in config which has a much higher budget ($0.43 per hour vs $0.26 per hour).
Hence request to change the instance creation configuration in Step 3 ( https://course.fast.ai/start_gcp.html#step-3-create-an-instance ) to the following:
export IMAGE_FAMILY="pytorch-latest-gpu" # or "pytorch-latest-cpu" for non-GPU instances export ZONE="us-central1-a" export INSTANCE_NAME="my-fastai-instance" export INSTANCE_TYPE="n1-highmem-8" # budget: "n1-highmem-4"
# budget: 'type=nvidia-tesla-k80,count=1' gcloud compute instances create $INSTANCE_NAME \ --zone=$ZONE \ --image-family=$IMAGE_FAMILY \ --image-project=deeplearning-platform-release \ --maintenance-policy=TERMINATE \ --accelerator="type=nvidia-tesla-p4,count=1" \ --machine-type=$INSTANCE_TYPE \ --boot-disk-size=200GB \ --metadata="install-nvidia-driver=True" \ --preemptible
In the fast ai setup guide for GCP, https://course.fast.ai/start_gcp.html
it was recommended to use us-west1-b. However Nvidia Tesla P4 is not available in this instance and is available in us-central1-a.
Further, in instance creation code, Nvidia tesla p100 was provided in config which has a much higher budget ($0.43 per hour vs $0.26 per hour).
Hence request to change the instance creation configuration in Step 3 ( https://course.fast.ai/start_gcp.html#step-3-create-an-instance ) to the following:
export IMAGE_FAMILY="pytorch-latest-gpu" # or "pytorch-latest-cpu" for non-GPU instances export ZONE="us-central1-a" export INSTANCE_NAME="my-fastai-instance" export INSTANCE_TYPE="n1-highmem-8" # budget: "n1-highmem-4"
# budget: 'type=nvidia-tesla-k80,count=1' gcloud compute instances create $INSTANCE_NAME \ --zone=$ZONE \ --image-family=$IMAGE_FAMILY \ --image-project=deeplearning-platform-release \ --maintenance-policy=TERMINATE \ --accelerator="type=nvidia-tesla-p4,count=1" \ --machine-type=$INSTANCE_TYPE \ --boot-disk-size=200GB \ --metadata="install-nvidia-driver=True" \ --preemptible