Fine-tuning jobs are unable to start. We are observing increased failure rates for new fine-tuning jobs. We are currently investigating for the root cause.

We have implemented a solution and the delayed fine-tune jobs are being processed. We are monitoring the progress.

The scope of the issue is limited to the finalizing step of new fine-tuning jobs which delays their completion after starting. Inference with fine-tuned models is unaffected. We continue to develop a solution