The Third Generation of Apple's Foundation Modelsmachinelearning.apple.com11 points·by 2bit·上個月·1 comments
ParaRNN: Large-Scale Nonlinear RNNs, Trainable in Parallelmachinelearning.apple.com2 points·by 2bit·3 個月前·0 comments
Exploring LLMs with MLX and the Neural Accelerators in the M5 GPUmachinelearning.apple.com2 points·by 2bit·8 個月前·0 comments
FastVLM: Efficient Vision Encoding for Vision Language Modelsmachinelearning.apple.com93 points·by 2bit·12 個月前·6 comments
Apple Intelligence Foundation Language Models Tech Report 2025machinelearning.apple.com242 points·by 2bit·12 個月前·204 comments
Updates to Apple's On-Device and Server Foundation Language Modelsmachinelearning.apple.com6 points·by 2bit·去年·1 comments
An LLM-Based Approach to Review Summarization on the App Storemachinelearning.apple.com4 points·by 2bit·去年·0 comments
Controlling Language and Diffusion Models by Transporting Activationsmachinelearning.apple.com90 points·by 2bit·去年·15 comments
2bit·去年·discussYou can find a similar method in zundo, a zustand middleware for undo/redo. It has the concept of pastStates and futureStates.