i'm pumped to announce dedicated container inference - one of the products i was so excited to build when i first joined @togethercompute and now 6 months later we're live! huge kudos to the team - this makes it easy to run inference for dense compute bound models such as video, audio, and avatar generation with primitives for auto-scaling, queueing, priorities, metrics, logging and more if you're building in this space, i would love to hear from you!