I have been trying to run a custom model training on vertex, but I keep getting an error due to vertex custom model training cpus limit, Im limited to 1 cpu which doesnt make sense because the minimum amount of cpus allowed for training is 4. I tried to increase the quota but I couldnt even submit a request and it said to contact sales support. I contacted them they keep putting me in a loop of things i tried and then they said that its something for another support team. Its been a week and I cant figure out what to do. My account isnt free trial I upgraded it 2 weeks ago and its paid.
Hi @Furix,
Welcome to Google Cloud Community!
From the information you’ve provided, it seems like you’re encountering an issue with custom model training on Vertex AI, specifically related to the CPU limit. The core problem is that you’re encountering a CPU limit (1 CPU) despite Vertex AI Custom Training’s minimum requirement of 4 CPUs, even though you’re on a paid account. Here’s a useful approach to help troubleshoot and resolve the issue:
Check Resource Quotas: In your Google Cloud Console, go to IAM & Admin > Quotas & System Limits and check the CPU quotas related to Vertex AI custom model training. Ensure that the quotas are sufficient and that there are no region-specific limits causing the 1-CPU restriction.
Compute Resources: Ensure that you are selecting the appropriate machine type (CPU, GPU, or TPU) for your custom model training job. If you’ve configured the job to use a minimum of 4 CPUs, ensure the machine type supports this and you’ve selected an appropriate configuration for the CPU count.
Contact Google Cloud Support: You mentioned that you tried to increase the quota but couldn’t even submit a request. Since you’ve already contacted the sales support team, it seems there is a more appropriate team that could handle your issue. In this case, I suggest contacting Google Cloud Support, as they can provide more insights and take a deeper look at the issue regarding CPU resource allocation problem. Provide them with detailed information about the problem, including error messages, steps you’ve taken, and the urgency of resolving the issue.
I hope the above information is helpful.