I want to deploy a machine learning model (more precisely stable diffusion) in the cloud and make it accessible via an API. I am relatively new to this topic, so I am not quite sure what services there are and what might best fit my requirements. This is just for a research topic, so I don't expect big loads and the response time is not crucial. More important would be the cost factor. I have already looked at implementing it with a VM with Cloud Engine but that seems to be rather expensive. Also, I don't plan to further train the model in the cloud, so it should be deployed once and then only be used for prediction. Obviously I need access to a GPU to render decent sized images. As already stated above, the model should be accessible via API, so the client sends a prompt from which a picture is generated. The model does not have to be running all the time, so it would be OK if there was some startup time.
It would be great if any of you could give me some tips or share their experiences with different services.
Thank you in advance!
