Step By Step/Collect Data

Collect Data

In this step you bring the data you'll train on into a dataset. Open your project and click Collect (step 2) in the project menu on the left.


1. Click Add Dataset

On the Datasets page, click Add Dataset.

The Datasets page with the Add Dataset button


2. Create the dataset

The Create Dataset dialog has two steps. First, give the dataset a Dataset Name (required) and, if you like, a Description, then click Next.

Create Dataset, step 1: name and description

Second, choose how your data is organized. The options depend on your project's task type, so follow the steps you see on screen. This example is an audio anomaly detection dataset: the data is split manually (you choose which files go to training and which to testing), and each label is inferred from the file name. Other task types may offer different choices.

Create Dataset, step 2: how data is split and how labels are inferred

Click Create Dataset to finish. Use Back to return to the first step, or Cancel to close the dialog.


3. Select your dataset

Your datasets are listed with how each one was set up (data acquisition, data split, labeling and labels inference). Selecting the right dataset is important: the Selected dataset is the one used by default when you set up training, so check it before you move on.

Open the ⋮ menu in the Actions column to manage a dataset:

The dataset list with the Actions menu open

  • Use Dataset — select it for your project.
  • Open — go to its upload page to add files.
  • Edit — change its details.
  • Delete — remove it.

4. Upload your files

Choose Open on your dataset to reach the Upload Files page. The page adapts to your task type, including the accepted file formats, so follow what you see on screen. In this audio anomaly detection example:

The Upload Files page

  • Which category? — pick where the files go. Here, Training (required) takes normal files only, Testing (required) takes a mix of normal and anomaly files, and Validation (optional) takes a mix of normal and anomaly files.
  • Files or Folder — upload individual files, or a whole folder.
  • Label hint — a note explains how labels are read. Here the label comes from the file name before the first underscore, so normal_001.wav is labeled normal and anomaly_001.wav is labeled anomaly.
  • Upload area — click to select your files. The accepted formats are listed at the top of the page (here .wav, .mp3, .ogg, .flac), along with the size limit (here 1 GB and 10,000 files per upload).

When the upload finishes, the page moves to the Done step.

Collect Data — Edge Studio Docs