ultralytics / hub

Ultralytics HUB tutorials and support
https://hub.ultralytics.com
GNU Affero General Public License v3.0
136 stars 13 forks source link

Dataset processing #11

Closed UnkoC0wbOy closed 2 years ago

UnkoC0wbOy commented 2 years ago

Search before asking

HUB Component

Datasets

Bug

Processing times are over 48 hours to upload relatively small datasets <500 images. Datasets have been downloaded straight from Roboflow using the YOLOv5 PyTorch option. I have had to manually change the .YAML file name to match the dataset and zipped file name. While trying to trouble shoot the issue I noticed that the downloaded file from Roboflow has a different file structure than the COCO6 dataset used as an example of what our dataset should look like. I understand that Ultralytics has no control over how the Roboflow datasets are configured....

Correcting the file structure would take manual file dragging and dropping while requiring edits to the .YAML file. If this is what is required I will just have to suck it up.. Perhaps I am missing something right in front of my face at the same time...

Environment

Ultralytics HUB Version v0.1.10-beta.2 Client User Agent Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/15.2 Safari/605.1.15 Operating System MacIntel Browser Window Size 1850 x 878 Server Timestamp 1641964834

Minimal Reproducible Example

No response

Additional

No response

kalenmike commented 2 years ago

@UnkoC0wbOy Thank you for submitting this bug. The dataset you are trying to upload is not formatted correctly, both the YAML and the folder structure of the zip are incorrect. I have corrected the format and attached the zip here for you to use as a reference.

Sheds.zip

I have also replaced your last attempt with the correctly formatted dataset which you should now be able to preview in your Ultralytics HUB account. The processing took less than 5 minutes if you would like to try again with the correctly formatted zip file.

We did identify a bug that prevented the dataset from moving to failed when submitted incorrectly. We will address this for the next release.

Let me know if there is anything else we can do to help.

kalenmike commented 2 years ago

Smaller datasets are now validated before upload (> v0.1.10-beta.3). An error message is displayed if the the dataset is formated incorrectly to help with debugging.

glenn-jocher commented 2 years ago

@kalenmike awesome, nice work!!

github-actions[bot] commented 2 years ago

👋 Hello, this issue has been automatically marked as stale because it has not had recent activity. Please note it will be closed if no further activity occurs.

Access additional YOLOv5 🚀 resources:

Access additional Ultralytics ⚡ resources:

Feel free to inform us of any other issues you discover or feature requests that come to mind in the future. Pull Requests (PRs) are also always welcomed!

Thank you for your contributions to YOLOv5 🚀 and Vision AI ⭐!