Why do it this way?

Quickest and dirtiest way to get going with Encord.

Typically importing LOCAL data using the UI is done when testing Encord’s capabilities or when you intend to import a relatively small amount of data, into Encord, for annotation.

If you only need Encord for annotation/annotation services and you know exactly what data you want to annotate, you might not need to use Index. This means importing your data to a mirrored Dataset, and then creating an Annotate Project.

Pros and Cons

ProsCons
  • Simplest import
  • Quicker Project set up
  • Lack of flexibility for data usage in Projects
  • No data management or curation
  • No cloud integration

Import Data

We’re going to import our dataset of fruit using the UI. The dataset is a mixture of images and videos of the following types of fruit:

  • Apples
  • Bananas
  • Blueberries
  • Cherries
  • Kiwi
  • Persimmons
  • Strawberries
1

Download Data

Download and extract the contents of Fruit-images-videos.zip file.

2

Create a Mirrored Dataset

Create the mirrored Dataset E2E Fruit Dataset using the UI. Using mirrored Datasets is the simplest way get data into Encord for annotation. Mirrored Datasets provide no method of curating or managing your data.

Mirrored datasets sync all the data to your Dataset automatically. This means if you want to add more data to your Dataset, you can add more data to the Dataset and then data automatically gets added to your Annotate Project.

3

Import Data

Import the data from Fruit-images-videos.zip to the mirrored Dataset E2E Fruit Dataset.

4

Re-encode Videos

The dataset we provided has some videos with audio and some other issues. We want to re-encode those videos with issues so our annotators have the best experience possible when labelling the videos.

We strongly recommend re-encoding any imported videos with issues. Re-encoding your videos ensures the best performance when annotating your data.

Was this page helpful?