Expand description
DFlash block-diffusion drafter (DFLASH-BRINGUP-PLAN.md, 2026-07-13).
5-layer qwen3-class mini-transformer that drafts a 16-token block in ONE non-causal
forward, conditioned on the TARGET’s hidden states at 6 tapped layers (concatenated
through fc + hidden_norm). No embed / lm_head of its own — the round reuses the
target’s. Reference: z-lab/dflash dflash/model.py (semantics frozen in the plan doc);
oracle: tools/dflash_oracle.py -> /data/cache/dflash-oracle.npz.
FIRST LIGHT = f32-resident weights + fresh full-context forward (no draft KV cache) — correctness vs the oracle, then the cache/quant/window arms land measurement-gated.
Structs§
- Dflash
Cfg - Dflash
Draft - Dflash
Kv - Draft KV cache (round-cost fix, 2026-07-13): per-layer normed+roped ctx K and raw ctx V, append-only in committed order. Block K/V land TRANSIENTLY at [len..len+b] each round (never committed — the reference crops them identically). Kills the per-round full-ctx projection recompute (first light was O(ctx)/round -> 7 tok/s).
- Dflash
Layer - Markov
Head