Files
ray/python/ray/util/sgd/torch/examples/sgd-development.yaml
T

99 lines
3.1 KiB
YAML

# An unique identifier for the head node and workers of this cluster.
cluster_name: sgd-pytorch
# The maximum number of workers nodes to launch in addition to the head
# node. This takes precedence over min_workers. min_workers default to 0.
min_workers: 2
initial_workers: 2
max_workers: 2
target_utilization_fraction: 0.9
# If a node is idle for this many minutes, it will be removed.
idle_timeout_minutes: 10
# docker:
# image: tensorflow/tensorflow:1.5.0-py3
# container_name: ray_docker
# Cloud-provider specific configuration.
provider:
type: aws
region: us-east-1
availability_zone: us-east-1c
# How Ray will authenticate with newly launched nodes.
auth:
ssh_user: ubuntu
# ssh_private_key: ...
head_node:
InstanceType: p3.2xlarge
ImageId: ami-0698bcaf8bd9ef56d
# KeyName: ...
InstanceMarketOptions:
MarketType: spot
SpotOptions:
BlockDurationMinutes: 360
BlockDeviceMappings:
- DeviceName: /dev/sda1
Ebs:
VolumeSize: 300
# SpotOptions:
# MaxPrice: "9.0"
worker_nodes:
InstanceType: p3.8xlarge
ImageId: ami-0698bcaf8bd9ef56d
# KeyName: ...
InstanceMarketOptions:
MarketType: spot
SpotOptions:
BlockDurationMinutes: 360
BlockDeviceMappings:
- DeviceName: /dev/sda1
Ebs:
VolumeSize: 300
# SpotOptions:
# MaxPrice: "9.0"
# # Run workers on spot by default. Comment this out to use on-demand.
# InstanceMarketOptions:
# MarketType: spot
setup_commands:
# This replaces the standard anaconda Ray installation
- ray || pip install -U https://s3-us-west-2.amazonaws.com/ray-wheels/latest/ray-0.9.0.dev0-cp36-cp36m-manylinux1_x86_64.whl
# Uncomment this and the filemount to update the Ray installation with your local Ray code
# - rm -rf ./anaconda3/lib/python3.6/site-packages/ray/util/sgd/
# - cp -rf ~/sgd ./anaconda3/lib/python3.6/site-packages/ray/util/
# Installing this without -U to make sure we don't replace the existing Ray installation
- pip install ray[rllib]
- pip install -U ipdb torch torchvision tqdm
# Install Apex
- rm -rf apex || true
- git clone https://github.com/NVIDIA/apex && cd apex && pip install -v --no-cache-dir ./ || true
file_mounts: {
# This should point to ray/python/ray/util/sgd.
# ~/sgd: ../../../sgd,
}
# Custom commands that will be run on the head node after common setup.
head_setup_commands: []
# Custom commands that will be run on worker nodes after common setup.
worker_setup_commands: []
# # Command to start ray on the head node. You don't need to change this.
head_start_ray_commands:
- ray stop
- ray start --head --object-manager-port=8076 --autoscaling-config=~/ray_bootstrap_config.yaml --object-store-memory=1000000000
# Command to start ray on worker nodes. You don't need to change this.
worker_start_ray_commands:
- ray stop
- ray start --address=$RAY_HEAD_IP:6379 --object-manager-port=8076 --object-store-memory=1000000000