---

license: apache-2.0

base_model: Qwen/Qwen3.6-35B-A3B

base_model_relation: quantized

pipeline_tag: text-generation

library_name: gguf

tags:

- gguf

- rocmfp4

- dflash

- speculative-decoding

- quantized

- imatrix

- amd

- rocm

- strix-halo

- moe

---

Qwen3.6-35B-A3B — ROCmFP4 STRIX_LEAN, DFlash baked in

A single-file, self-accelerating GGUF: the model and its DFlash speculative-decoding draft are merged into one .gguf. No --model-draft, no --spec-type flag — point -m at this file and speculative decoding just happens.

To our knowledge, the first "draft-included" GGUF publication anywhere.

llama-server -m Qwen3.6-35B-A3B-STRIX_LEAN-DFLASH.gguf -ngl 999 -fa on --jinja -c 65536

Requirements

This needs both a ROCmFP4-aware build and the DFlash-graft support for embedded drafts — neither exists upstream yet. Use:

• gsrunion/rocmfp4-llama branch dflash-graft (built and validated on AMD Strix Halo / gfx1151), or
• your own build once the fixes below land in charlie12345/rocmfp4-llama

On first load the server extracts the draft's tensors to a small cached sidecar file next to the model (one-time, ~1 second).

Measured performance

AMD Ryzen AI Max+ 395 (Strix Halo, 128 GB unified LPDDR5X), server-timing, self-accelerating load (zero extra flags):

| | tok/s | acceptance |

|---|---|---|

| Baked single-file | 91.8 | 98.5% (405/411) |

| Two-file (--model-draft + flags) | 96.0 | 97–98% |

| Plain LEAN, no draft | 63.1 | — |

Within noise of the two-file config — the merge adds no overhead.

How it was made

The draft's tensors are merged into the target GGUF prefixed dflash.* (target keeps its own tensor names untouched — no collision, no size overhead: DFlash drafts already borrow the target's token embeddings and output head at runtime, so nothing is duplicated). A dflash.embedded marker key flags the file for auto-detection.

Two fixes were needed in the serving fork to make this work (both filed against the base fork, worth watching if you hit similar issues building your own):

1. The tensor-count sanity check in the model loader didn't allow "extra" tensors belonging to a sibling model in the same file — even though the check already had unused plumbing for exactly this case.
2. The draft's mask_token_id (namespaced under tokenizer.* by convention, though it's actually draft-specific) has to be copied into the merged file explicitly, or drafting silently no-ops with zero speedup and no error.

Base weights: gsrunion/Qwen3.6-35B-A3B-ROCmFP4-STRIX_LEAN-GGUF. Draft: z-lab/Qwen3.6-35B-A3B-DFlash.

Credits

• Base model: Qwen — Qwen3.6-35B-A3B (Apache-2.0)
• DFlash draft: z-lab
• ROCmFP4 quant formats: Hal0ai; fork base: charlie12345