[sgd] Example for Training (#5292)

2026-08-02 13:01:01 +08:00 · 2019-07-27 01:10:25 -07:00
parent 06fec63c87
commit 7e715520e5
4 changed files with 116 additions and 0 deletions
@@ -0,0 +1,73 @@
+# 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: 1
+initial_workers: 1
+max_workers: 1
+
+target_utilization_fraction: 0.9
+
+# If a node is idle for this many minutes, it will be removed.
+idle_timeout_minutes: 20
+# 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-1f
+
+# How Ray will authenticate with newly launched nodes.
+auth:
+    ssh_user: ubuntu
+
+head_node:
+    InstanceType: p3.8xlarge
+    ImageId: ami-0757fc5a639fe7666
+    #    InstanceMarketOptions:
+    #        MarketType: spot
+    #        SpotOptions:
+    #            MaxPrice: "9.0"
+
+
+worker_nodes:
+    InstanceType: p3.8xlarge
+    ImageId: ami-0757fc5a639fe7666
+    # InstanceMarketOptions:
+    #     MarketType: spot
+        # SpotOptions:
+        #     MaxPrice: "9.0"
+
+    #     # Run workers on spot by default. Comment this out to use on-demand.
+    #     InstanceMarketOptions:
+    #         MarketType: spot
+
+setup_commands:
+    - ray || pip install -U https://s3-us-west-2.amazonaws.com/ray-wheels/latest/ray-0.8.0.dev2-cp36-cp36m-manylinux1_x86_64.whl
+    - conda install -y pytorch torchvision cudatoolkit=9.0 -c pytorch
+    - pip install -U ipdb ray[rllib]
+
+
+file_mounts: {
+}
+
+# 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 --redis-port=6379 --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 --redis-address=$RAY_HEAD_IP:6379 --object-manager-port=8076 --object-store-memory=1000000000
+