vishnukool

Good question, I've heard other engineers ask for this. The MVP version doesn't yet handle multiple instances, but it will be on our roadmap.

We built a tool called spotML to make training on AWS/GCP cheaper.

Spot Instances are 70% cheaper than On-Demand instances but are prone to interruptions. We mitigate the downside of these interruptions through the use of persistence features, including optional fallback to On-Demand instances. So you can optimize workflows according to your budget and time constraints.

History: We were working on a neural rendering startup that needed a lot of GAN training which was getting very expensive. We were blowing roughly $1000, to train a single category class. Training on Spot instances was cheaper, but still a mess. It needed lot of hand holding/devops stuff to make it usable. So we built SpotML to automate a lot of things.

Posting it here to see if the community finds this helpful, so that we can open it up to the larger community.

does it need to be an app ? The friction of downloading an app is discouraging to try it out.

Thanks for trying it out ! Ah, yes. It doesn't work great with sectionals yet. The model does an Ok job on the kind of sofa styles it's seen during training but still needs to generalize better. Also the UI demo is a quick sub second inference. But the quality gets progressively better with iterative refinement via the API which isn't happening right now.

Thanks, do let me know any feedback when you try it. Regarding the name, yeah we'll probably need to pick something more generic. Right now though it's only home decor categories like sofa, chairs but the idea is to expand to broader categories.

Thanks, yes. The 3D asset that's generated is pretty low poly at the moment. But I see how this could be used as a base for further modeling work.