CLI import not uploading annotations (only uploads images)

This category is for question related to accounts on https://app.roboflow.com/

Please share the following so we may better assist you:

  1. Project type (Object detection, Classification, Polygon, etc.); Object Detection
  2. The operating system & browser you are using and their versions; Ubuntu, CLI
  3. The screenshot of the error being triggered in your browser’s developer tools/console. Please copy/paste the url below to watch how to pull up your devtools

I am attempting to upload a dataset to Roboflow with the roboflow import cli command.

I downloaded a dataset from Universe with the following:

mkdir data-with-labels; cd data-with-labels; curl -L “https://universe.roboflow.com/ds/YZbicMV8Z4?key=WrbJD7E7Ky” > roboflow.zip; unzip roboflow.zip; rm roboflow.zip; cd …

Then, I attempted to import the dataset with:

roboflow auth
roboflow import data-with-labels/

This command uploaded the images, but not the annotations. Is there a way to upload annotations with the CLI?

Hi @Robert_Shaw, you’ll need to update your command for upload. The command you’ve used will only upload images.

Mohamed,

My dataset in the the YOLOv5 format, so the directory looks like:

data-with-labels/data.yaml
data-with-labels/train/images
data-with-labels/train/labels
data-with-labels/valid/images
data-with-labels/valid/labels
data-with-labels/test/images
data-with-labels/test/labels

My understanding is that the roboflow import command is designed to be used in this situation - is there any documentation on how to use the import command to do upload a dataset structured in the YOLOv5 format?

I just downloaded one of my datasets to run the upload flow. I’ll create a video example of how it works.

Hi @Robert_Shaw, I just attempted this with my own projects, and found the same issue you presented.

I’ve prepared and submitted a full bug report for this issue to our Product Team.

I’ve marked the bug as well so I can keep you up to date on the status.

Hello @Mohamed, any update on this?

It is working now - I actually just tested it again to confirm