Apple MPS transformer decoding shows abrupt latency spikes up to 21x in narrow decoding-budget intervals due to KV cache and execution regime shifts, absent on CPU and CUDA.
Title resolution pending
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
verdicts
UNVERDICTED 3representative citing papers
Prompt tuning matches full model tuning performance on large language models while tuning only a small fraction of parameters and improves robustness to domain shifts.
SGD is reformulated via a master equation from discrete updates, producing a discrete Fokker-Planck equation that predicts non-stationary variance growth proportional to learning rate in flat Hessian directions.
citing papers explorer
-
Non-Monotonic Latency in Apple MPS Decoding: KV Cache Interactions and Execution Regimes
Apple MPS transformer decoding shows abrupt latency spikes up to 21x in narrow decoding-budget intervals due to KV cache and execution regime shifts, absent on CPU and CUDA.
-
The Power of Scale for Parameter-Efficient Prompt Tuning
Prompt tuning matches full model tuning performance on large language models while tuning only a small fraction of parameters and improves robustness to domain shifts.
-
Why SGD is not Brownian Motion: A New Perspective on Stochastic Dynamics
SGD is reformulated via a master equation from discrete updates, producing a discrete Fokker-Planck equation that predicts non-stationary variance growth proportional to learning rate in flat Hessian directions.