Understanding Inputs & Outputs

A Gradient Workflow is composed of a series of steps. These steps specify how to orchestrate computational tasks. Each step can communicate with other steps through what are known as inputs and outputs.

There are three types of inputs and outputs. Understanding how these function will help you craft concise and elegant Workflows.

  • Datasets

  • Volumes

  • Strings

Datasets

The dataset type leverages the Gradient platform native dataset primitive. Information stored within datasets is not limited to any single type of data. In fact, a generic dataset can include anything from pretrained models to generated images to configuration files. Inherent to datasets is the notion of versions. Workflows can consume and produce new dataset versions as well as tag new versions of existing datasets. Note: datasets must be defined in advance of being referenced in a workflow. See the dataset documentation for more info.

Scenario 1: Consuming a dataset that already exists within Gradient

inputs:
my-dataset:
type: dataset
with:
id: my-dataset-id

Scenario 2: Generating a new dataset version from a Workflow step

my-job:
with:
args:
- bash
- '-c'
- cp -R /my-trained-model /outputs/my-dataset
image: bash:5
outputs:
my-dataset:
type: dataset
with:
id: my-dataset-id

Volumes

Unlike, e.g. GitHub Actions, it is likely that multiple Gradient Steps/Actions will execute on multiple compute nodes. To facilitate the passing of data between these nodes, Gradient Actions exposes the notion of volumes and volume passing.

Volumes enable actions such as the @git-checkout action. Volumes can be defined as input volumes or output volumes or both. When a volume is an output it is mounted in /outputs and is writeable. When a volume is an input it is mounted in /inputs and is read only.

outputs:
my-volume:
type: volume

In this example a volume is first created as an output and then used as an input in a subsequent job step:

defaults:
resources:
instance-type: P4000
jobs:
job1:
with:
args:
- bash
- -c
- echo hello > /outputs/my-volume/testfile1; echo "wrote testfile1 to volume"
image: bash
outputs:
my-volume:
type: volume
job2:
needs:
- job1
with:
args:
- bash
- -c
- cat /inputs/my-volume/testfile1
image: bash
inputs:
my-volume: job1.outputs.my-volume

Volumes cannot currently be used as an output after the job they were created with, this limitation is planned to be removed in the future.

Strings

In some cases, you may need to pass a single value between Workflow steps. The string type makes this possible.

Scenario 1: Passing a string from as a workflow-level input

inputs:
my-string:
type: string
with:
value: "my string value"
jobs:
job-1:
resources:
instance-type: P4000
with:
args:
- bash
- -c
- cat /inputs/my-string
image: bash:5
inputs:
my-string: workflow.inputs.my-string

Scenario 2: Passing a string between job steps

defaults:
resources:
instance-type: P4000
jobs:
job-1:
with:
args:
- bash
- -c
- echo "string output from job-1" > /outputs/my-string; echo job-1 finished
image: bash:5
outputs:
my-string:
type: string
job-2:
with:
args:
- bash
- -c
- cat /inputs/my-string
image: bash:5
needs:
- job-1
inputs:
my-string: job-1.outputs.my-string

Scenario 3: Creating a model from a dataset and passing the model ID as a string to a Deployment step

NOTE: There is no native Gradient Actions for Model Deployments today. Instead, you can use the Gradient SDK to create and manage your inference endpoints.

To run this example you will need to a) create a dataset named test-model and upload valid tensorflow model files to it, b) define a secret named MY_API_KEY with your gradient-cli api-key, c) substitute your clusterId in the deployment create step.

defaults:
resources:
instance-type: P4000
jobs:
UploadModel:
uses: create-[email protected]
with:
name: my-model
type: Tensorflow
inputs:
model:
type: dataset
with:
id: test-model
outputs:
model-id:
type: string
DeployModel:
needs:
- UploadModel
inputs:
model-id: UploadModel.outputs.model-id
env:
PAPERSPACE_API_KEY: secret:MY_API_KEY
with:
command: bash
args:
- -c
- >-
gradient deployments create
--clusterId cl1234567
--deploymentType TFServing
--modelId $(cat inputs/model-id)
--name "Sample Deployment"
--machineType P4000
--imageUrl tensorflow/serving:latest-gpu
--instanceCount 1
image: paperspace/gradient-sdk