[Projects] Unify hyphen vs underscore handling for arguments (#6208)

This commit is contained in:
Philipp Moritz 2019-11-20 23:52:41 -08:00 committed by GitHub
parent 1f9ab74293
commit a4437813eb
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
6 changed files with 5 additions and 86 deletions

View file

@ -13,7 +13,7 @@ environment:
commands:
- name: run-sync
command: python sync_parameter_server.py --num-workers {{num_workers}}
command: python sync_parameter_server.py --num-workers {{num-workers}}
help: "Start the synchronous parameter server."
params:
- name: num-workers
@ -24,7 +24,7 @@ commands:
tmux: true
- name: run-async
command: python async_parameter_server.py --num-workers {{num_workers}}
command: python async_parameter_server.py --num-workers {{num-workers}}
help: "Start the asynchronous parameter server."
params:
- name: num-workers

View file

@ -1,22 +0,0 @@
# This file is generated by `ray project create`.
# A unique identifier for the head node and workers of this cluster.
cluster_name: ray-example-resnet
# The maximum number of workers nodes to launch in addition to the head
# node. This takes precedence over min_workers. min_workers defaults to 0.
max_workers: 1
# Cloud-provider specific configuration.
provider:
type: aws
region: us-west-2
availability_zone: us-west-2a
head_node:
InstanceType: m5.2xlarge
ImageId: ami-0b294f219d14e6a82 # Deep Learning AMI (Ubuntu) Version 21.0
# How Ray will authenticate with newly launched nodes.
auth:
ssh_user: ubuntu

View file

@ -1,58 +0,0 @@
# This file is generated by `ray project create`.
name: ray-example-resnet
description: "Using ray to train resnet on multiple gpus"
tags: ["ray-example", "machine-learning", "tensorflow", "resnet"]
documentation: https://ray.readthedocs.io/en/latest/auto_examples/plot_resnet.html
cluster: .rayproject/cluster.yaml
environment:
requirements: .rayproject/requirements.txt
commands:
- name: train
command: |
if [ "{{dataset}}" == "cifar10" ]; then
# Get the CIFAR-10 dataset.
curl -o cifar-10-binary.tar.gz https://www.cs.toronto.edu/~kriz/cifar-10-binary.tar.gz;
tar -xvf cifar-10-binary.tar.gz;
else
# Get the CIFAR-100 dataset.
curl -o cifar-100-binary.tar.gz https://www.cs.toronto.edu/~kriz/cifar-100-binary.tar.gz;
tar -xvf cifar-100-binary.tar.gz;
fi
python resnet_main.py --dataset {{dataset}} --train_data_path {{train_data_path}} --eval_data_path {{eval_data_path}} --eval_dir {{eval_data_path}} --eval_batch_count {{eval_batch_count}} --num_gpus {{num_gpus}}
params:
- name: dataset
help: "The dataset to train on."
default: "cifar10"
choices: ["cifar10", "cifar100"]
- name: train-data-path
help: "Data path for the training data."
default: "'cifar-10-batches-bin/data_batch*'"
type: str
- name: eval-data-path
help: "Data path for the testing data."
default: "cifar-10-batches-bin/test_batch.bin"
type: str
- name: eval-dir
help: "Data path for the tensorboard logs."
default: "/tmp/resnet-model/eval"
type: str
- name: eval-batch-count
help: "Number of batches to evaluate over."
default: 50
type: int
- name: num-gpus
help: "Number of GPUs to use for training."
default: 0
type: int
config:
tmux: true
output_files: [
# Save the logs from the latest run in snapshots.
"/tmp/ray/session_latest/logs"
]

View file

@ -1 +0,0 @@
ray[rllib,debug]

View file

@ -12,7 +12,7 @@ environment:
commands:
- name: run
command: python streaming.py --num-mappers {{num_mappers}} --num-reducers {{num_reducers}}
command: python streaming.py --num-mappers {{num-mappers}} --num-reducers {{num-reducers}}
help: "Start the streaming example."
params:
- name: num-mappers

View file

@ -105,9 +105,9 @@ class ProjectDefinition:
"Parameter {} has type {} which is not supported. "
"Type must be one of {}".format(
name, param["type"], list(types.keys())))
parser.add_argument("--" + name, **param)
parser.add_argument("--" + name, dest=name, **param)
parsed_args = parser.parse_args(list(args)).__dict__
parsed_args = vars(parser.parse_args(list(args)))
if wildcards:
for key, val in parsed_args.items():