Data sources

A data source connects a project to data that lives somewhere else, such as a cloud storage bucket, an organizational dataset, or an upload portal, and keeps pulling from it on a schedule. A data source is a pipeline: an ordered list of steps where the first step fetches data into the project, and the steps after it act on the project once the new data has landed. New data can arrive, retrain the model, and build a fresh deployment without anyone opening the Studio.

What happens on each run

A pipeline run has three steps:

The import step fetches data from the source you configured, labeling each sample according to the strategy you chose.
The post-sync actions run against the project, in the order they appear in the pipeline. These are ordinary Studio operations, so retraining here is the same retraining you would trigger by hand.
The run is reported, by email and optionally to a webhook, including how many new items arrived.

Keep in mind:

Building a deployment requires retraining. Enable the retrain action as well, because otherwise there is no new model to build from.
Retraining requires the project to have been trained at least once, since the pipeline reuses the existing impulse rather than designing one.
A run that imports nothing still runs the actions you enabled. Check the new item count in the notification to tell an empty run from a broken one.

For the first run, set Prefix to a small folder in your bucket rather than the whole bucket. Labels are assigned at import time, so if the labeling strategy is wrong, only a few samples need fixing. Once the labels look right, widen the path in the pipeline configuration.

Run transformation jobs directly from your projects

You can also trigger cloud jobs, known as transformation blocks, these are particularly useful if you want to generate synthetic datasets or automate tasks using the Edge Impulse API. We provide several pre-built transformation blocks available for organizations’ projects:

Data sources overview showing a successful pipeline connection — Data sources

This view, originally accessible from the main left menu, has been moved to the Data acquisition tab for better clarity. The screenshots have not yet been updated.

Add a data source

Click in + Add new data source and select where your data lives: You can either use:

Cloud data storage
Organizational datasets (enterprise feature)
Upload portals (enterprise feature)
Transformation blocks (enterprise feature)
Don’t import data (if you just need to create a pipeline)

Add new data source dialog with storage provider choices — Add new data source

See how to configure your storage buckets:

Click on Next, provide credentials:

Data source credentials step with bucket and authentication fields — Provide your credentials

Click on Verify credentials:

Automatic labeling step for configuring label inference from folders or filenames — Automatically label your data

Here, you have several options to automatically label your data:

Infer from folder name

In the example above, the structure of the folder is the following:

.
├── cars
│   ├── cars.01741.jpg
│   ├── cars.01743.jpg
│   ├── cars.01745.jpg
│   ├── ... (400 items)
├── unknown
│   ├── unknown.test_2547.jpg
│   ├── unknown.test_2548.jpg
│   ├── unknown.test_2549.jpg
│   ├── ... (400 items)
└── unlabeled
    ├── cars.02066.jpg
    ├── cars.02067.jpg
    ├── cars.02068.jpg
    └── ... (14 items)

3 directories, 814 files

The labels will be picked from the folder name and will be split between your training and testing set using the following ratio 80/20.

The samples present in an unlabeled/ folder will be kept unlabeled in Edge Impulse Studio.

Alternatively, you can also organize your folder using the following structure to automatically split your dataset between training and testing sets:

.
├── testing
│   ├── cars
│   │   ├── cars.00012.jpg
│   │   ├── cars.00031.jpg
│   │   ├── cars.00035.jpg
│   │   └── ... (~150 items)
│   └── unknown
│       ├── unknown.test_1012.jpg
│       ├── unknown.test_1026.jpg
│       ├── unknown.test_1027.jpg
│       ├── ... (~150 items)
├── training
│   ├── cars
│   │   ├── cars.00006.jpg
│   │   ├── cars.00025.jpg
│   │   ├── cars.00065.jpg
│   │   └── ... (~600 items)
│   └── unknown
│       ├── unknown.test_1002.jpg
│       ├── unknown.test_1005.jpg
│       └── unknown.test_46.jpg
│       └── ... (~600 items)
└── unlabeled
    ├── cars.02066.jpg
    ├── cars.02067.jpg
    ├── cars.02068.jpg
    └── ... (14 items)

7 directories, 1512 files

Infer from file name

When using this option, only the file name is taken into account. The part before the first . will be used to set the label. E.g. cars.01741.jpg will set the label to cars.

Keep the data unlabeled

All the data samples will be unlabeled, you will need to label them manually before using them. Finally, click on Next, post-sync actions.

Trigger actions step for selecting post-ingestion processing actions — Trigger actions

From this view, you can automate several actions:

Recreate data explorer The data explorer gives you a one-look view of your dataset, letting you quickly label unknown data. If you enable this you’ll also get an email with a screenshot of the data explorer whenever there’s new data.
Retrain model If needed, will retrain your model with the same impulse. If you enable this you’ll also get an email with the new validation and test set accuracy. Note: You will need to have trained your project at least once.
Create new version Store all data, configuration, intermediate results and final models.
Create new deployment Builds a new library or binary with your updated model. Requires ‘Retrain model’ to also be enabled.

Run the pipeline

Once your pipeline is set, you can run it directly from the UI, from external sources or by scheduling the task.

Run pipeline panel with controls for starting an ingestion job — Run your pipeline

Run the pipeline from the UI

To run your pipeline from Edge Impulse studio, click on the ⋮ button and select Run pipeline now.

Run the pipeline from code

To run your pipeline from Edge Impulse studio, click on the ⋮ button and select Run pipeline from code. This will display an overlay with curl, Node.js and Python code samples.

You will need to create an API key to run the pipeline from code.

Run pipeline from code panel with command-line instructions — Run the pipeline from code

Schedule your pipeline jobs

By default, your pipeline will run every day. To schedule your pipeline jobs, click on the ⋮ button and select Edit pipeline.

Free users can only run the pipeline every 4 hours. If you are an enterprise customer, you can run this pipeline up to every minute.

Pipeline editor with schedule and notification settings for a data source — Edit pipeline

Once the pipeline has successfully finish, you will receive an email like the following:

Email notification example containing full data source pipeline results — Email example containing the full results

You can also define who can receive the email. The users must be collaborators on your project. See Collaboration.

Webhooks

You can also create a webhook to call a URL when the pipeline has run. It will run a POST request containing the following information:

{
    "organizationId":XX,
    "pipelineId":XX,
    "pipelineName":"Import data from portal \"Data sources demo\"",
    "projectId":XXXXX,
    "success":true,
    "newItems":0,
    "newChecklistOK":0,
    "newChecklistFail":0
}

success reports whether the run completed and newItems the number of samples added. Several runs in a row with zero new items can mean a broken credential or a wrong path. The checklist counters report how many incoming items passed or failed the dataset checklist, for sources that have one.

Data source webhooks panel with endpoint and event settings — Data sources webhooks

Edit your pipeline

To update your pipeline, edit the configuration JSON available in ⋮ -> Run pipeline from code. Here is an example of what you can get if all the actions have been selected:

[
    {
        "name": "Fetch data from s3://data-pipeline/data-pipeline-example/infer-from-folder/",
        "builtinTransformationBlock": {
            "type": "s3-to-project",
            "endpoint": "https://s3.your-endpoint.com",
            "path": "s3://data-pipeline/data-pipeline-example/infer-from-folder/",
            "region": "fr-par",
            "accessKey": "XXXXX",
            "category": "split",
            "labelStrategy": "infer-from-folder-name",
            "secretKeyEncrypted": "xxxxxx"
        }
    },
    {
        "name": "Refresh data explorer",
        "builtinTransformationBlock": {
            "type": "project-action",
            "refreshDataExplorer": true
        }
    },
    {
        "name": "Retrain model",
        "builtinTransformationBlock": {
            "type": "project-action",
            "retrainModel": true
        }
    },
    {
        "name": "Create new version",
        "builtinTransformationBlock": {
            "type": "project-action",
            "createVersion": true
        }
    },
    {
        "name": "Create on-device deployment (C++ library)",
        "builtinTransformationBlock": {
            "type": "project-action",
            "buildBinary": "zip",
            "buildBinaryModelType": "int8"
        }
    }
]

Free projects have only access to the above builtinTransformationBlock. If you are part of an organization, you can use your custom transformation jobs in the pipeline. In your organization workspace, go to Custom blocks -> Transformation and select Run job on the job you want to add.

Transformation blocks list used to select pipeline processing steps — Transformation blocks

Select Copy as pipeline step and paste it to the configuration json file.

Copy action for adding a transformation block command to a data source pipeline — Copy

Additional resources

Data pipelines for the organization-level equivalent that can feed datasets rather than projects
Cloud data storage for configuring bucket credentials
Transformation blocks for writing a custom step
Versioning for what the Create new version action stores

OVERVIEW

PROJECTS

ORGANIZATIONS

What happens on each run

Run transformation jobs directly from your projects

Add a data source

Infer from folder name

Infer from file name

Keep the data unlabeled

Run the pipeline

Run the pipeline from the UI

Run the pipeline from code

Schedule your pipeline jobs

Webhooks

Edit your pipeline

Additional resources

​What happens on each run

​Run transformation jobs directly from your projects

​Add a data source

​Infer from folder name

​Infer from file name

​Keep the data unlabeled

​Run the pipeline

​Run the pipeline from the UI

​Run the pipeline from code

​Schedule your pipeline jobs

​Webhooks

​Edit your pipeline

​Additional resources

What happens on each run

Run transformation jobs directly from your projects

Add a data source

Infer from folder name

Infer from file name

Keep the data unlabeled

Run the pipeline

Run the pipeline from the UI

Run the pipeline from code

Schedule your pipeline jobs

Webhooks

Edit your pipeline

Additional resources