Experiment Template REST API
caution
Note: The Experiment API is in the alpha stage which is subjected to incompatible changes in future releases.
Developers can register a parameterized experiment as an experiment template, For example, if the developer wants to change the following "--learning_rate=0.1" to parameters.
"experimentSpec": {
"meta": {
"cmd": "python /var/tf_mnist/mnist_with_summaries.py --log_dir=/train/log --learning_rate=0.1 --batch_size=150"
},
"...": "..."
}
They can use two curly braces as placeholders, the template format will be as
"experimentSpec": {
"meta": {
"cmd": "python /var/tf_mnist/mnist_with_summaries.py --log_dir=/train/log --learning_rate={{learning_rate}} --batch_size=150"
},
"...": "..."
}
The template parameters format will be as
{
"name": "learning_rate",
"value": 0.1,
"required": true,
"description": "This is learning_rate of training."
}
name: placeholder name value; default value required: Indicates whether the user must enter parameters, when required is true, value can be null description: Introduction of this parameter
Users can use existing experiment templates and adjust the default value to create experiments. After the user submits the experiment template, the submarine server finds the corresponding template based on the name. And the template handler converts input parameters to an actual experiment, such as a distributed TF experiment.
The "replicas", "cpu", "memory" of resources will be automatically parameterized, so developers do not need to add them. For example, if there are "Ps" and "Worker" under spec, the following parameters will be automatically appended.
spec.Ps.replicas
spec.Ps.resourceMap.cpu
spec.Ps.resourceMap.memory
spec.Worker.replicas
spec.Worker.resourceMap.cpu
spec.Worker.resourceMap.memory
Create experiment templateβ
POST /api/v1/template
Parametersβ
Field Name | Type | In | Description |
---|---|---|---|
name | String | body | Experiment template name. This is required. |
author | String | body | Author name. |
description | String | body | Description of the experiment template. |
parameters | List<ExperimentTemplateParamSpec> | body | Parameters of the experiment template. |
experimentSpec | ExperimentSpec | body | Spec of the experiment template. |
ExperimentTemplateParamSpecβ
Field Name | Type | Description |
---|---|---|
name | String | Parameter name. |
required | Boolean | true / false. Whether the parameter is required. |
description | String | Description of the parameter. |
value | String | Value of the parameter. |
ExperimentSpecβ
Field Name | Type | Description |
---|---|---|
meta | ExperimentMeta | Meta data of the experiment template. |
environment | EnvironmentSpec | Environment of the experiment template. |
spec | Map<String, ExperimentTaskSpec> | Spec of pods. |
code | CodeSpec | Experiment codespec. |
ExperimentMetaβ
Field Name | Type | Description |
---|---|---|
name | String | Experiment Name. |
namespace | String | Experiment namespace. |
framework | String | Experiment framework. |
cmd | String | Command. |
envVars | Map<String, String> | Environmental variables. |
EnvironmentSpecβ
See more details in environment api.
ExperimentTaskSpecβ
Field Name | Type | Description |
---|---|---|
replicas | Integer | Numbers of replicas. |
resoureces | String | Resouces of the task |
name | String | Task name. |
image | String | Image name. |
cmd | String | Command. |
envVars | Map<String, String> | Environmental variables. |
CodeSpecβ
Field Name | Type | Description |
---|---|---|
syncMode | String | sync mode of code spec. |
url | String | url of code spec. |
Exampleβ
Example Request
curl -X POST -H "Content-Type: application/json" -d '
{
"name": "my-tf-mnist-template",
"author": "author",
"description": "This is a template to run tf-mnist",
"parameters": [{
"name": "learning_rate",
"value": 0.1,
"required": true,
"description": "This is learning_rate of training."
},
{
"name": "batch_size",
"value": 150,
"required": true,
"description": "This is batch_size of training."
},
{
"name": "experiment_name",
"value": "tf-mnist1",
"required": true,
"description": "the name of experiment."
}
],
"experimentSpec": {
"meta": {
"cmd": "python /var/tf_mnist/mnist_with_summaries.py --log_dir=/train/log --learning_rate={{learning_rate}} --batch_size={{batch_size}}",
"name": "{{experiment_name}}",
"envVars": {
"ENV1": "ENV1"
},
"framework": "TensorFlow",
"namespace": "default"
},
"spec": {
"Ps": {
"replicas": 1,
"resources": "cpu=1,memory=1024M"
},
"Worker": {
"replicas": 1,
"resources": "cpu=1,memory=1024M"
}
},
"environment": {
"image": "apache/submarine:tf-mnist-with-summaries-1.0"
}
}
}
' http://127.0.0.1:32080/api/v1/template
Example Response
{
"status":"OK",
"code":200,
"success":true,
"message":null,
"result":{
"experimentTemplateId":{
"id":1,
"serverTimestamp":1626160071451
},
"experimentTemplateSpec":{
"name":"my-tf-mnist-template",
"author":"author",
"description":"This is a template to run tf-mnist",
"parameters":
[{
"name":"learning_rate",
"required":"true",
"description":"This is learning_rate of training.",
"value":"0.1"
},
{
"name":"batch_size",
"required":"true",
"description":"This is batch_size of training.",
"value":"150"
},
{
"name":"experiment_name",
"required":"true",
"description":"the name of experiment.",
"value":"tf-mnist1"
},
{
"name":"spec.Ps.replicas",
"required":"false",
"description":"",
"value":"1"
},
{
"name":"spec.Ps.resourceMap.cpu",
"required":"false",
"description":"",
"value":"1"
},
{
"name":"spec.Ps.resourceMap.memory",
"required":"false",
"description":"",
"value":"1024M"
},
{
"name":"spec.Worker.replicas",
"required":"false",
"description":"",
"value":"1"
},
{
"name":"spec.Worker.resourceMap.cpu",
"required":"false",
"description":"",
"value":"1"
},
{
"name":"spec.Worker.resourceMap.memory",
"required":"false",
"description":"","
value":"1024M"
}],
"experimentSpec":{
"meta":{
"name":"{{experiment_name}}",
"namespace":"default",
"framework":"TensorFlow",
"cmd":"python /var/tf_mnist/mnist_with_summaries.py --log_dir\u003d/train/log --learning_rate\u003d{{learning_rate}} --batch_size\u003d{{batch_size}}",
"envVars":{"ENV1":"ENV1"}
},
"environment":{
"name":null,
"dockerImage":null,
"kernelSpec":null,
"description":null,
"image":"apache/submarine:tf-mnist-with-summaries-1.0"
},
"spec":{
"Ps":{
"replicas":1,
"resources":"cpu\u003d1,memory\u003d1024M",
"name":null,
"image":null,
"cmd":null,
"envVars":null,
"resourceMap":{
"memory":"1024M",
"cpu":"1"
}
},
"Worker":{
"replicas":1,
"resources":"cpu\u003d1,memory\u003d1024M",
"name":null,
"image":null,
"cmd":null,
"envVars":null,
"resourceMap":{
"memory":"1024M",
"cpu":"1"
}
}
},
"code":null
}
}
},
"attributes":{}
}
List experiment templateβ
GET /api/v1/template
Exampleβ
Example Request
curl -X GET http://127.0.0.1:32080/api/v1/template
Example Response
{
"status":"OK",
"code":200,
"success":true,
"message":null,
"result":{
[{
"experimentTemplateId":{
"id":1,
"serverTimestamp":1626160071451
},
"experimentTemplateSpec":{
"name":"my-tf-mnist-template",
"author":"author",
"description":"This is a template to run tf-mnist",
"parameters":
[{
"name":"learning_rate",
"required":"true",
"description":"This is learning_rate of training.",
"value":"0.1"
},
{
"name":"batch_size",
"required":"true",
"description":"This is batch_size of training.",
"value":"150"
},
{
"name":"experiment_name",
"required":"true",
"description":"the name of experiment.",
"value":"tf-mnist1"
},
{
"name":"spec.Ps.replicas",
"required":"false",
"description":"",
"value":"1"
},
{
"name":"spec.Ps.resourceMap.cpu",
"required":"false",
"description":"",
"value":"1"
},
{
"name":"spec.Ps.resourceMap.memory",
"required":"false",
"description":"",
"value":"1024M"
},
{
"name":"spec.Worker.replicas",
"required":"false",
"description":"",
"value":"1"
},
{
"name":"spec.Worker.resourceMap.cpu",
"required":"false",
"description":"",
"value":"1"
},
{
"name":"spec.Worker.resourceMap.memory",
"required":"false",
"description":"","
value":"1024M"
}],
"experimentSpec":{
"meta":{
"name":"{{experiment_name}}",
"namespace":"default",
"framework":"TensorFlow",
"cmd":"python /var/tf_mnist/mnist_with_summaries.py --log_dir\u003d/train/log --learning_rate\u003d{{learning_rate}} --batch_size\u003d{{batch_size}}",
"envVars":{"ENV1":"ENV1"}
},
"environment":{
"name":null,
"dockerImage":null,
"kernelSpec":null,
"description":null,
"image":"apache/submarine:tf-mnist-with-summaries-1.0"
},
"spec":{
"Ps":{
"replicas":1,
"resources":"cpu\u003d1,memory\u003d1024M",
"name":null,
"image":null,
"cmd":null,
"envVars":null,
"resourceMap":{
"memory":"1024M",
"cpu":"1"
}
},
"Worker":{
"replicas":1,
"resources":"cpu\u003d1,memory\u003d1024M",
"name":null,
"image":null,
"cmd":null,
"envVars":null,
"resourceMap":{
"memory":"1024M",
"cpu":"1"
}
}
},
"code":null
}
}
}],
"attributes":{}
}
Get experiment templateβ
GET /api/v1/template/{name}
Parametersβ
Field Name | Type | In | Description | Required |
---|---|---|---|---|
name | String | path | Experiment template name. | o |
Exampleβ
Example Request
curl -X GET http://127.0.0.1:32080/api/v1/template/my-tf-mnist-template
Example Response
{
"status":"OK",
"code":200,
"success":true,
"message":null,
"result":{
"experimentTemplateId":{
"id":1,
"serverTimestamp":1650788898882
},
"experimentTemplateSpec":{
"name":"my-tf-mnist-template",
"author":"author",
"description":"This is a template to run tf-mnist",
"parameters":[
{
"name":"learning_rate",
"required":"true",
"description":"This is learning_rate of training.",
"value":"0.1"
},
{
"name":"batch_size",
"required":"true",
"description":"This is batch_size of training.",
"value":"150"
},
{
"name":"experiment_name",
"required":"true",
"description":"the name of experiment.",
"value":"tf-mnist1"
},
{
"name":"spec.Ps.replicas",
"required":"false",
"description":"",
"value":"1"
},
{
"name":"spec.Ps.resourceMap.cpu",
"required":"false",
"description":"",
"value":"1"
},
{
"name":"spec.Ps.resourceMap.memory",
"required":"false",
"description":"",
"value":"1024M"
},
{
"name":"spec.Worker.replicas",
"required":"false",
"description":"",
"value":"1"
},
{
"name":"spec.Worker.resourceMap.cpu",
"required":"false",
"description":"",
"value":"1"
},
{
"name":"spec.Worker.resourceMap.memory",
"required":"false",
"description":"",
"value":"1024M"
}
],
"experimentSpec":{
"meta":{
"experimentId":null,
"name":"{{experiment_name}}",
"namespace":"default",
"framework":"TensorFlow",
"cmd":"python /var/tf_mnist/mnist_with_summaries.py --log_dir\u003d/train/log --learning_rate\u003d{{learning_rate}} --batch_size\u003d{{batch_size}}",
"envVars":{
"ENV1":"ENV1"
},
"tags":[]
},
"environment":{
"name":null,
"dockerImage":null,
"kernelSpec":null,
"description":null,
"image":"apache/submarine:tf-mnist-with-summaries-1.0"
},
"spec":{
"Ps":{
"replicas":1,
"resources":"cpu\u003d1,memory\u003d1024M",
"name":null,
"image":null,
"cmd":null,
"envVars":null,
"resourceMap":{
"memory":"1024M",
"cpu":"1"
}
},
"Worker":{
"replicas":1,
"resources":"cpu\u003d1,memory\u003d1024M",
"name":null,
"image":null,
"cmd":null,
"envVars":null,
"resourceMap":{
"memory":"1024M",
"cpu":"1"
}
}
},
"code":null
}
}
},
"attributes":{}
}
Patch templateβ
PATCH /api/v1/template/{name}
Parametersβ
Field Name | Type | In | Description | Required |
---|---|---|---|---|
name | String | path and body | Experiment template name. | o |
author | String | body | Author name. | o |
description | String | body | Description of the experiment template. | x |
parameters | List<ExperimentTemplateParamSpec> | body | Parameters of the experiment template. | o |
experimentSpec | ExperimentSpec | body | Spec of the experiment template. | o |
Exampleβ
Example Request
curl -X PATCH -H "Content-Type: application/json" -d '
{
"name": "my-tf-mnist-template",
"author": "author-new",
"description": "This is a template to run tf-mnist",
"parameters": [{
"name": "learning_rate",
"value": 0.1,
"required": true,
"description": "This is learning_rate of training."
},
{
"name": "batch_size",
"value": 150,
"required": true,
"description": "This is batch_size of training."
},
{
"name": "experiment_name",
"value": "tf-mnist1",
"required": true,
"description": "the name of experiment."
}
],
"experimentSpec": {
"meta": {
"cmd": "python /var/tf_mnist/mnist_with_summaries.py --log_dir=/train/log --learning_rate={{learning_rate}} --batch_size={{batch_size}}",
"name": "{{experiment_name}}",
"envVars": {
"ENV1": "ENV1"
},
"framework": "TensorFlow",
"namespace": "default"
},
"spec": {
"Ps": {
"replicas": 1,
"resources": "cpu=1,memory=1024M"
},
"Worker": {
"replicas": 1,
"resources": "cpu=1,memory=1024M"
}
},
"environment": {
"image": "apache/submarine:tf-mnist-with-summaries-1.0"
}
}
}
' http://127.0.0.1:32080/api/v1/template/my-tf-mnist-template
Example Response
{
"status":"OK",
"code":200,
"success":true,
"message":null,
"result":{
"experimentTemplateId":{
"id":2,
"serverTimestamp":1626160071451
},
"experimentTemplateSpec":{
"name":"my-tf-mnist-template",
"author":"author-new",
"description":"This is a template to run tf-mnist",
"parameters":
[{
"name":"learning_rate",
"required":"true",
"description":"This is learning_rate of training.",
"value":"0.1"
},
{
"name":"batch_size",
"required":"true",
"description":"This is batch_size of training.",
"value":"150"
},
{
"name":"experiment_name",
"required":"true",
"description":"the name of experiment.",
"value":"tf-mnist1"
},
{
"name":"spec.Ps.replicas",
"required":"false",
"description":"",
"value":"1"
},
{
"name":"spec.Ps.resourceMap.cpu",
"required":"false",
"description":"",
"value":"1"
},
{
"name":"spec.Ps.resourceMap.memory",
"required":"false",
"description":"",
"value":"1024M"
},
{
"name":"spec.Worker.replicas",
"required":"false",
"description":"",
"value":"1"
},
{
"name":"spec.Worker.resourceMap.cpu",
"required":"false",
"description":"",
"value":"1"
},
{
"name":"spec.Worker.resourceMap.memory",
"required":"false",
"description":"",
"value":"1024M"
}],
"experimentSpec":{
"meta":{
"name":"{{experiment_name}}",
"namespace":"default",
"framework":"TensorFlow",
"cmd":"python /var/tf_mnist/mnist_with_summaries.py --log_dir\u003d/train/log --learning_rate\u003d{{learning_rate}} --batch_size\u003d{{batch_size}}",
"envVars":{"ENV1":"ENV1"}
},
"environment":{
"name":null,
"dockerImage":null,
"kernelSpec":null,
"description":null,
"image":"apache/submarine:tf-mnist-with-summaries-1.0"
},
"spec":{
"Ps":{
"replicas":1,
"resources":"cpu\u003d1,memory\u003d1024M",
"name":null,
"image":null,
"cmd":null,
"envVars":null,
"resourceMap":{"memory":"1024M","cpu":"1"}
},
"Worker":{
"replicas":1,
"resources":"cpu\u003d1,memory\u003d1024M",
"name":null,
"image":null,
"cmd":null,
"envVars":null,
"resourceMap":{"memory":"1024M","cpu":"1"}
}
},
"code":null
}
}
},
"attributes":{}
}
note
"description", "parameters", "experimentSpec", "author" etc can be updated using this API. "name" of experiment template is not supported.
Delete templateβ
GET /api/v1/template/{name}
Parametersβ
Field Name | Type | In | Description | Required |
---|---|---|---|---|
name | String | path | Experiment template name. | o |
Exampleβ
Example Request
curl -X DELETE http://127.0.0.1:32080/api/v1/template/my-tf-mnist-template
Example Response
{
"status":"OK",
"code":200,
"success":true,
"message":null,
"result":{
"experimentTemplateId":{
"id":2,
"serverTimestamp":1626160071451
},
"experimentTemplateSpec":{
"name":"my-tf-mnist-template",
"author":"author-new",
"description":"This is a template to run tf-mnist",
"parameters":
[{
"name":"learning_rate",
"required":"true",
"description":"This is learning_rate of training.",
"value":"0.1"
},
{
"name":"batch_size",
"required":"true",
"description":"This is batch_size of training.",
"value":"150"
},
{
"name":"experiment_name",
"required":"true",
"description":"the name of experiment.",
"value":"tf-mnist1"
},
{
"name":"spec.Ps.replicas",
"required":"false",
"description":"",
"value":"1"
},
{
"name":"spec.Ps.resourceMap.cpu",
"required":"false",
"description":"",
"value":"1"
},
{
"name":"spec.Ps.resourceMap.memory",
"required":"false",
"description":"",
"value":"1024M"
},
{
"name":"spec.Worker.replicas",
"required":"false",
"description":"",
"value":"1"
},
{
"name":"spec.Worker.resourceMap.cpu",
"required":"false",
"description":"",
"value":"1"
},
{
"name":"spec.Worker.resourceMap.memory",
"required":"false",
"description":"",
"value":"1024M"
}],
"experimentSpec":{
"meta":{
"name":"{{experiment_name}}",
"namespace":"default",
"framework":"TensorFlow",
"cmd":"python /var/tf_mnist/mnist_with_summaries.py --log_dir\u003d/train/log --learning_rate\u003d{{learning_rate}} --batch_size\u003d{{batch_size}}",
"envVars":{"ENV1":"ENV1"}
},
"environment":{
"name":null,
"dockerImage":null,
"kernelSpec":null,
"description":null,
"image":"apache/submarine:tf-mnist-with-summaries-1.0"
},
"spec":{
"Ps":{
"replicas":1,
"resources":"cpu\u003d1,memory\u003d1024M",
"name":null,
"image":null,
"cmd":null,
"envVars":null,
"resourceMap":{"memory":"1024M","cpu":"1"}
},
"Worker":{
"replicas":1,
"resources":"cpu\u003d1,memory\u003d1024M",
"name":null,
"image":null,
"cmd":null,
"envVars":null,
"resourceMap":{"memory":"1024M","cpu":"1"}
}
},
"code":null
}
}
},
"attributes":{}
}
Use template to create a experimentβ
POST /api/v1/experiment/{template_name}
Parametersβ
Field Name | Type | In | Description | Required |
---|---|---|---|---|
name | String | path and body | Experiment template name. | o |
params | Map<String, String> | body | Parameters of the experiment including experiment_name . | o |
Exampleβ
Example Request
curl -X POST -H "Content-Type: application/json" -d '
{
"name": "tf-mnist",
"params": {
"learning_rate":"0.01",
"batch_size":"150",
"experiment_name":"newexperiment1"
}
}
' http://127.0.0.1:32080/api/v1/experiment/my-tf-mnist-template
Example Response
{
"status":"OK",
"code":200,
"success":true,
"message":null,
"result":{
"experimentId":"experiment_1626160071451_0001",
"name":"newexperiment1",
"uid":"b895985c-411c-4e89-90e0-c60a2a8a4235",
"status":"Accepted",
"acceptedTime":"2021-07-13T16:21:31.000+08:00",
"createdTime":null,
"runningTime":null,
"finishedTime":null,
"spec":{
"meta":{
"name":"newexperiment1",
"namespace":"default",
"framework":"TensorFlow",
"cmd":"python /var/tf_mnist/mnist_with_summaries.py --log_dir\u003d/train/log --learning_rate\u003d0.01 --batch_size\u003d150",
"envVars":{"ENV1":"ENV1"}
},
"environment":{
"name":null,
"dockerImage":null,
"kernelSpec":null,
"description":null,
"image":"apache/submarine:tf-mnist-with-summaries-1.0"
},
"spec":{
"Ps":{
"replicas":1,
"resources":"cpu\u003d1,memory\u003d1024M",
"name":null,
"image":null,
"cmd":null,
"envVars":null,
"resourceMap":{"memory":"1024M","cpu":"1"}
},
"Worker":{
"replicas":1,
"resources":"cpu\u003d1,memory\u003d1024M",
"name":null,
"image":null,
"cmd":null,
"envVars":null,
"resourceMap":{"memory":"1024M","cpu":"1"}
}
},
"code":null
}
},
"attributes":{}
}