AVTensor: A High-Performance Rust Media Decoder for Training Pipelinesrunwayml.com1 points·by nielka·yesterday·0 comments
Borrowing the Night: Reclaiming Idle Inference GPUs for Researchrunwayml.com2 points·by nielka·9 days ago·0 comments
Why Distributed Training Is Hard: DTensor and the Costs of Abstractionrunwayml.com3 points·by nielka·2 months ago·0 comments
Stop Writing YAML: Configuring ML Systems with Confingyrunwayml.com4 points·by nielka·2 months ago·0 comments
60x Faster Cold Starts: Treating Peer GPUs as Weight Serversrunwayml.com5 points·by nielka·2 months ago·0 comments
No Idle GPUs: Managing Research Compute at Runwayrunwayml.com6 points·by nielka·3 months ago·0 comments
Reboot: Rebuild civilization by reinventing lost technologiesksindi.com41 points·by nielka·4 months ago·4 comments