From the source
Lead story
Ask your AI
Top stories
Models & availability
Latest
Trajectory as the Teacher: Few-Step Discrete Flow Matching via Energy-Navigated Distillation
Apple researchers propose Trajectory-Shaped Discrete Flow Matching (TS-DFM), a method that improves few-step discrete flow matching by using a lightweight energy compass to guide token selection during training.
On a 170M-parameter language model, the 8-step student achieves 32% lower perplexity than the 1,024-step teacher while being 128× faster.
From the source
Trajectory-Shaped Discrete Flow Matching (TS-DFM) replaces these blind jumps with guided navigation: a lightweight energy compass evaluates candidate continuations at each midpoint, selecting the most coherent.
machinelearning.apple.com
