1 réponse
- Le plus récent
- Le plus de votes
- La plupart des commentaires
0
The concurrency of a real-time endpoint depends on the number of workers maintained inside your algorithm container. For each worker, a copy of the model weights need to be loaded. In other words, we need to first configure the container to maintain multiple workers and make sure there is enough CPU & GPU memory to host multiple models.
I think for stable diffusion, the officially recommended GPU memory is 10 GB and g4dn.xlarge only comes with 16, which is not sufficient for 2 models running concurrently ?
Could you please check the runtime GPU utilization as well as the configured number of workers in your container ?
répondu il y a un an
Contenus pertinents
- demandé il y a un an
- demandé il y a 4 mois
- demandé il y a 7 mois
- demandé il y a un an
- AWS OFFICIELA mis à jour il y a 2 ans
- AWS OFFICIELA mis à jour il y a un an
- AWS OFFICIELA mis à jour il y a 2 ans
thanks, i'll check that is it possible concurrently process requests by using other instance type having more GPU memory.