My dev machines are old and can’t handle Axon training runs. I’d like to use rent-by-the-hour GPU enabled servers from an outfit like Linode, Digital Ocean, GCP, Azure, AWS, etc.
What provider is best? Any tips on optimal configuration re: Ram, CPU, GPU?
Thanks in Advance!
Worth checking out GCP Vertex AI. Create a model using custom training | Vertex AI | Google Cloud
They have a nice api where you can run one off training jobs with gpus. Just need to spec a docker container.
It’s worth noting that you can prototype on GPUs with Colab for free: