Roboflow Serverless Inference: A Thousand Models on a Shared GPU Fleetblog.roboflow.com·2 pts·yeldarb·0
Show HN: NYCerebro, semantic search of NYC traffic cams (written by v0)nycerebro.vercel.app·4 pts·yeldarb·2
Clip Meets Model Zoo Experts: Pseudo-Supervision for Visual Enhancementmachinelearning.apple.com·4 pts·yeldarb·0
NnScaler converts models designed for one GPU to run concurrently on multiplegithub.com·5 pts·yeldarb·0
Searchformer: Beyond A* – Better planning with transformers via search dynamicsgithub.com·181 pts·yeldarb·28
HuggingFace JAT: Distributed online training of a general multitask RL Agentgithub.com·4 pts·yeldarb·0
Show HN: Autodistill – Use big slow foundation models to train small fast modelsgithub.com·18 pts·yeldarb·0