The default steps on the website for serverless create broken containers that I am charged for.

Hi! I'm new to all of this, but I'm running into a serious issue. I've put money on my account, click "Get Started" on autoscale with serverless, click "Text" under preset models, then click the deepseek-ai/DeepSeek-R1 model. There are no prompts to configure anything - everything is made with RunPod's premade settings. I click the unmodified "Hello World" in the web prompt, and it puts it into a delay queue while it spins up workers that hang before going into an unhealthy state, and then crashing. I've spent about a dollar just waiting for the hello world prompt to finish, and I have no idea what's happening. I haven't touched or configured anything at all - I'm just following the recommended, in-service steps. What is happening?
3 Replies
nerdylive
nerdylive4w ago
maybe check the logs?, deepseek R1 is so heavy that i believe no current setup in runpod can run the whole model it requires multiple high-end GPU's
riverfog7
riverfog72w ago
8xMI300X should work with 1.5 terabytes of vram
nerdylive
nerdylive2w ago
Yup But serverless doesn't allow that I believe

Did you find this page helpful?