Ultralytics Cloud Training Starts but Shows No Progress and Then Disconnects

Support Ultraltyics Platform
1

2b121e5e-4dc7-49c2-8f91-f79b2374714c
2b121e5e-4dc7-49c2-8f91-f79b2374714c1600×730 99.9 KB

f8df6e95-c62e-4f62-ab23-c6385cc8e3ca
f8df6e95-c62e-4f62-ab23-c6385cc8e3ca1600×671 56.3 KB

Dear Ultralytics Support Team,

I am encountering an issue with Ultralytics Cloud training.

The training job behavior is as follows:

  1. I click “Start Training”

  2. The job initializes successfully

  3. No training progress or logs appear

  4. After some time, the status changes to Disconnected

  5. Final progress remains at 0%

This behavior is consistent across multiple models.

Configuration details:

  • Model: YOLOv8n / YOLOv11n

  • Task: Detection

  • Epochs: 100

  • Sufficient account balance available

  • Training initiated via Ultralytics Cloud

Could you please review the backend logs for my account and advise on the root cause?

Thank you for your assistance.

2

I would suggest trying out Ultralytics Platform

It’s more stable. We can arrange for migration of your HUB credits once you create an account.

4

Thanks for the update. We have migrated your credits to Platform. You can try training now. You can download your dataset from HUB and upload it to Platform ss zip and it should be imported with labels.

5

Noted and thank you.

6

Glad it’s sorted. Your credits should now be available in Ultralytics Platform—if you try a new Ultralytics YOLO run and it ever sits at 0% again or flips to Disconnected, the quickest checks are that the dataset imported in ready status (has train + val/test images and at least one labeled image) and then just retry the job; the Platform Cloud Training page has the exact “stuck at 0%” troubleshooting notes in the docs on Cloud Training.

Also, for privacy, I’d recommend editing your earlier post to remove the email address since this is a public thread. If anything still looks off, share the Platform model run link or run name here and we can take a look.