Manually Converting Large Files to WebM
  • Dark
    Light
  • PDF

Manually Converting Large Files to WebM

  • Dark
    Light
  • PDF

As explained in our guide on WebM and Frame-Accurate Annotation, video files over 1.07 GB (1,073,741,824 bytes) are not automatically converted to WebM format when uploaded to the Dataloop platform. Therefore, we recommend that you manually convert these files to WebM format in order to ensure frame-accurate annotation.

Converting Files to WebM

  1. Go to FaaS > Application Hub.
  2. Make sure only the box By Dataloop is checked, find custom-webm-converter under Library, and click Install.

    image.png

  3. Edit the Installation Configuration settings if needed (you can read more about the FaaS User-Interface).
  4. A new service will be deployed in your project and can be used in two ways:
    a. In your Dataset Browser, select the item you wish to convert to WebM, click the DATALOOP custom-webm-converter function under Application, and click Execute.

    image.png


    b. Create a new pipeline with a FaaS component, choose the run function from the new service you deployed, and start the pipeline. For complete information on pipelines, click here.

    image.png

    You can execute the webm-converter function by adding a trigger to the FaaS component in the pipeline or by executing the pipeline for a specific item in the Dataset Browser:

    image.png

  5. View the Item Metadata and Functions for information about the execution of the webm-converter.

    image.png

    The item’s conversion to webm is updated in the item’s metadata:
 "custom-webm-converter-8482855": {
 	"webm_module": {
 		"run": {
 			"620289f5a070aafbee1b11d1": {
 				"status": "success",
 				"progress": 100
 			}
 		}
 	}
 }
{
	"ref": "62028a9969edce1d21c83f8f",
	"type": "replace",
	"timestamp": "1970-01-20T00:45:33.721Z",
	"name": "61e034e40303c37bc345a6f4.webm",
	"refType": "id",
	"stream": "https://gate.dataloop.ai/api/v1/items/62028a9969edce1d21c83f8f/stream"
}