license: other
license_name: qwen-community-1.0
base_model:

• orcarouter/Qwen3.8-Flash-Next-Uncensored
• Qwen/Qwen3.8-Flash-Next
  base_model_relation: quantized
  pipeline_tag: text-generation
  library_name: gguf
  tags:
• gguf
• rocmfp4
• imatrix
• qwen4exp
• llama.cpp
• strix-halo
• gfx1151
• rocm
• amd
• ryzen-ai-max
• uncensored
• research

🔧 Runtime: build the ROCmFPX fork below

Stock llama.cpp will not load this file. You need both the qwen4exp architecture
and the ROCmFP4 tensor types in one tree. Our fork
kingjones30/ROCmFPX (fork of
charlie12345/ROCmFPX, branch main) has both.

git clone https://github.com/kingjones30/ROCmFPX.git
cd ROCmFPX
cmake -B build -DGGML_HIP=ON -DGPU_TARGETS=gfx1151 -DGGML_NATIVE=ON -DCMAKE_BUILD_TYPE=Release
cmake --build build --target llama-server llama-quantize -j$(nproc)

Qwen3.8-Flash-Next-Uncensored — ROCmFP4 STRIX_LEAN imatrix GGUF — AMD Ryzen AI Max+ 395 / gfx1151

The importance-matrix-calibrated STRIX_LEAN build — the highest-quality tier in this
family. STRIX_LEAN already spends more bits than FAST (Q5 token embeddings and PLE, half the
attention at higher precision); importance-weighted quantization on top gives it the lowest
perplexity of any tier here.

⚠️ Research artifact. Refusal behaviour has been removed. This does not add capability — it
removes guardrails. Use it deliberately, in a context where that is appropriate, and own the output.

Measured quality — held-out WikiText-2 raw, -c 512

4.9865 is the lowest perplexity across the whole family — −6.7% vs plain FAST, and −0.9% below
the FAST imatrix tier at the same calibration. That gap over FAST is the richer STRIX_LEAN recipe
(higher-precision embeddings/PLE) plus the imatrix weighting.

Honesty note: I do not publish a "plain STRIX_LEAN vs imatrix STRIX_LEAN" number, because
the source BF16 was reclaimed after the build and I can't re-quantize a non-imatrix STRIX_LEAN
for a clean same-recipe delta. The clean isolated imatrix effect (same recipe, imatrix on/off) is
the FAST tier's −5.9%; expect STRIX_LEAN's isolated imatrix gain to be in the same range.
imatrix moves quality, not speed — decode t/s is unchanged.

• Calibration corpus: bartowski
  calibration_datav3.
• ⚠️ imatrix computed on the 4-bit model (the 51.2B PLE + 128 GB GTT ceiling blocks a BF16
  forward pass on Strix Halo).

Speculative decoding (MTP)

draft-mtp works on this arch once you apply qwen4exp-mtp-graph.patch
(bundled) — it fixes the graph combiner that otherwise held acceptance near 0.36. Pair with the
stock Flash-Next MTP head from
kingjones777/Qwen3.8-Flash-Next-MTP-Heads-GGUF.
Measured acceptance on the FAST tier with the fixed graph was 0.94 (+27.7% tok/s); STRIX_LEAN
uses the same MTP head and the same graph. The head is stock Flash-Next — it only proposes drafts,
the main model verifies every token, so it never alters this model's output.

llama-server -m Qwen3.8-Flash-Next-Uncensored-Q4_0-ROCmFP4-STRIX_LEAN-imatrix-00001-of-00003.gguf \
  -md mtp-Qwen3.8-Flash-Next-Q6_K.gguf --spec-type draft-mtp \
  --spec-draft-n-min 2 --spec-draft-n-max 4 --n-gpu-layers-draft 99 \
  -ngl 999 -fa on -np 1 -c 131072 --jinja

⛔ ngram-mod at ≥64K can wedge the GPU

Fine at 32K; at ≥64K a context-checkpoint restore desyncs the QSA indexer cache and wedges the
SDMA queue (hard power-off to recover). Workaround (no new binary): -ctxcp 0 -cpent -1.
draft-mtp is unaffected. Field report:
…-STRIX-GGUF#6.

Recipe

Quantized from the BF16 weights published by
orcarouter/Qwen3.8-Flash-Next-Uncensored
— the abliteration is theirs. 4.78 bpw, 98.5 GiB.

Building the runtime

Two patches, both bundled: qwen4exp-on-rocmfpx-d3ca537.patch (arch, 156 KB) and
qwen4exp-mtp-graph.patch (draft-mtp fix).

git clone https://github.com/charlie12345/ROCmFPX.git
cd ROCmFPX && git checkout d3ca537
curl -LO https://huggingface.co/kingjones777/Qwen3.8-Flash-Next-Uncensored-ROCmFP4-STRIX_LEAN-imatrix-GGUF/resolve/main/qwen4exp-on-rocmfpx-d3ca537.patch
git apply qwen4exp-on-rocmfpx-d3ca537.patch
curl -LO https://huggingface.co/kingjones777/Qwen3.8-Flash-Next-Uncensored-ROCmFP4-STRIX_LEAN-imatrix-GGUF/resolve/main/qwen4exp-mtp-graph.patch   # optional
git apply qwen4exp-mtp-graph.patch
cmake -B build -DGGML_HIP=ON -DGPU_TARGETS=gfx1151 -DGGML_NATIVE=ON -DCMAKE_BUILD_TYPE=Release
cmake --build build --target llama-server llama-quantize -j$(nproc)

Speed — Ryzen AI MAX+ 395, gfx1151, ROCm 7.2.4, full offload

Decode speed is the same as the plain STRIX_LEAN build (same layout): 23.20 tok/s gen /
377.8 tok/s prompt / 63.3 GiB GTT, one fixed 6,963-token prompt, cache_prompt:false, median
of 4. Native max context 262,144 on a 128 GB box.

Files

Usage

llama-server \
  --model Qwen3.8-Flash-Next-Uncensored-Q4_0-ROCmFP4-STRIX_LEAN-imatrix-00001-of-00003.gguf \
  --mmproj mmproj-Qwen3.8-Flash-Next-Uncensored-BF16.gguf \
  --host 127.0.0.1 --port 8080 \
  --n-gpu-layers 999 --flash-attn on --fit off \
  --ctx-size 131072 --threads 16 --jinja

Do not use --no-mmap (and do not use -dio). The PLE table streams from the file through the
page cache; forcing it into anonymous memory gets the process OOM-killed with nothing in the log.

Reproduction

quantize: llama-quantize --imatrix unc.imatrix <BF16> <out> Q4_0_ROCMFP4_STRIX_LEAN 16
ppl     : llama-perplexity -m <this> -f wiki.test.raw -ngl 999 -fa on -dev ROCm0 -c 512   (NO -dio)

A number without its binary is a rumour — every figure above is measured on the fork runtime above.

Acknowledgements

charlie12345/ROCmFPX — ROCmFP4 tensor formats (MIT).
llama.cpp — engine, GGUF, conversion tooling.
AMD ROCm — ROCm 7.2.4, gfx1151. orcarouter — the
uncensored BF16 checkpoint. Qwen team — the base model. License qwen-community-1.0.