We just made LLM inference up to 30% faster β‘
Introducing DNeo, a new diffusion drafter designed in collaboration with @cerebras researchers.
The secret? Your draft model needs a working memory, and the "right" context encoding mechanism.
π Paper: https://t.co/Pj1qpeqpO9
We just made LLM inference up to 30% faster β‘
Introducing DNeo, a new diffusion drafter designed in collaboration with @cerebras researchers.
The secret? Your draft model needs a working memory, and the "right" context encoding mechanism.
π Paper: https://t.co/Pj1qpeqpO9
6/6 β The Next Generation of Diffusion!
Diffusion language models can be powerful, if you architect them correctly. DNeo provides a strong blueprint for dLLMs.
π» Code: https://t.co/CBXSYzLjAv
π€ Weights: https://t.co/IbHYYsHlg3
Questions + comments welcome!
5/6 β The Fused DNeo Head
Parallel drafting struggles with left-to-right consistency.
We combine DSpark's full-vocabulary correction with DFlash-2's context-aware scoring in a shared low-rank space.
The result: More expressive corrections.