NVIDIA has released Gated DeltaNet-2, a linear attention layer that decouples erasing old content from writing new content via separate channel-wise gates. At 1.3B parameters trained on 100B tokens, it outperforms Mamba-2, Gated DeltaNet and KDA on language modelling and long-context retrieval. https://www.marktechpost.com/2026/05/24/nvidia-ai-releases-gated-deltanet-2-a-linear-attention-layer-that-decouples-erase-and-write-in-the-delta-rule/ #AIagent #AI #GenAI #ML
Related
AI hedge fund Situational Awareness may have sold its public portfolio, but it still has its Anthropic shares | TechCrun...
AI hedge fund Situational Awareness may have sold its public portfolio, but it still has its Anthropic shares | TechCrunchThe former OpenAI researcher’s fund was forced to unwind p...
Jak dopadla čtvrtečnà obchodnà seance na burzách v Evropě a ve Spojených státech, jak se v druhém čtvrtletà dařilo ekono...
Jak dopadla čtvrtečnà obchodnà seance na burzách v Evropě a ve Spojených státech, jak se v druhém čtvrtletà dařilo ekonomikám #USA, eurozóny a #ČR a může #AI přispět k proměně afri...
How to look better on webcam — tips for the average person to look better and more professionalSome tips for the average...
How to look better on webcam — tips for the average person to look better and more professionalSome tips for the average non-streaming webcam user to look better and more professio...