Run AI inference apps with self-hosted models on Cloud Run with Nvidia GPUscloud.google.com 15 pointsLyalinDotCom2 years ago2 commentsSaveHideCopy link On HNComments−wietsevenema2yOne NVIDIA L4 GPU (24GB vRAM) per Cloud Run instance (many instances per Cloud Run service).Scale to zero: When there are no incoming requests, Cloud Run stops all remaining instances and you’re not charged.Fast cold start: When scaling from zero, processes in the container can use the GPU in approximately 5 seconds.Open large language models up to 13B parameters run great, including: Gemma 2 (9B), Llama 3.1 (8B), Mistral (7B), Qwen2 (7B).You can get Gemma 2 (2B, Q4_0) to return tokens after 11 seconds from a cold start (best case).−steren2yCloud Run PM here, ask me anything!
Comments
One NVIDIA L4 GPU (24GB vRAM) per Cloud Run instance (many instances per Cloud Run service).
Scale to zero: When there are no incoming requests, Cloud Run stops all remaining instances and you’re not charged.
Fast cold start: When scaling from zero, processes in the container can use the GPU in approximately 5 seconds.
Open large language models up to 13B parameters run great, including: Gemma 2 (9B), Llama 3.1 (8B), Mistral (7B), Qwen2 (7B).
You can get Gemma 2 (2B, Q4_0) to return tokens after 11 seconds from a cold start (best case).
Cloud Run PM here, ask me anything!