DeepSeek-V4-Flash-0731-ROCmFP4

Quants in this repo

Both target AMD gfx1151 (Ryzen AI MAX+ 395 / Strix Halo) and require a llama.cpp with the Q4_0_ROCMFP4_* quant types — see charlie12345/ROCmFPX.

⚠️ Note on the STRIX (105) tier: it protects attention K/V but not the LM head. On large-vocabulary models we now prefer tier 102 COHERENT plus explicit --output-tensor-type q6_K --token-embedding-type q6_K. This STRIX build predates that finding.

DeepSeek-V4-Flash-0731 — ROCmFP4 (Strix Halo) GGUF

This is a ROCmFP4 quant of deepseek-ai/DeepSeek-V4-Flash-0731, built to fit a single AMD Strix Halo box (128 GB unified memory) with full GPU offload. As far as I can tell it's the first ROCmFP4 quant of this model. I made it with the ROCmFPX fork of llama.cpp for the gfx1151 (Radeon 8060S / Ryzen AI MAX+ 395) Vulkan/ROCm stack.

Why I made it

A standard 4-bit GGUF of this model comes out around 141 GB, which overflows a 128 GB Strix Halo's shared pool and spills to CPU. I wanted the largest-quality quant that still fully offloads on a single box and stays coherent, so I mixed the expert tensors down to land it at ~101 GB.

Recipe (quantized from the F16 with the fork's llama-quantize):

• base type Q2_0_ROCMFPX
• ffn_down_exps → q3_0_rocmfpx (3.5 bpw)
• ffn_gate_exps, ffn_up_exps → q2_0_rocmfpx (2.5 bpw)
• attention / embeddings → ROCmFPX; norms kept in fp32

The ROCmFP4 (_ROCMFPX) types hold quality better than equivalent-bit k-quants on this hardware while using the FP4 paths on gfx1151.

Running it

Build the ROCmFPX fork (llama-server / llama-cli) for gfx1151, then:

export HSA_OVERRIDE_GFX_VERSION=11.5.1 GGML_HIP_ENABLE_UNIFIED_MEMORY=1
export AMD_VULKAN_ICD=RADV VK_ICD_FILENAMES=/usr/share/vulkan/icd.d/radeon_icd.json

./llama-server \
  -m DeepSeek-V4-Flash-0731-Q3-ROCmFP4-00001-of-00004.gguf \
  -dev Vulkan0 -ngl 999 -fa on -fit off --no-mmap \
  -c 8192 -n 2048 -np 1 -b 1024 -ub 512 -t 16 --poll 50 --jinja \
  --reasoning-format deepseek \
  --chat-template-kwargs '{"enable_thinking":false}' \
  --host 0.0.0.0 --port 8084

Notes from getting it stable on my box:

• -fit off — the fork's auto-fit step crashed on this arch for me; pin -ngl 999 and turn it off.
• --no-mmap — important for MoE speed. With mmap, experts page-fault per token and throughput roughly halves.
• -c 8192 with -b 1024 -ub 512 keeps the graph pool under its limit; larger context can overflow it.
• -n 2048 caps runaway generations so one request can't hold the single slot forever.
• --chat-template-kwargs '{"enable_thinking":false}' gives fast, direct answers. Drop it (or pass enable_thinking:true per request) for the model's reasoning mode.
• Expect roughly 5–8 tok/s — it's a 101 GB model on one iGPU. Use streaming for a usable feel.

A note on MTP

This checkpoint ships a multi-token-prediction (nextn) head, and I kept those tensors in this quant. I got a working MTP inference path running on this arch and tested it thoroughly, but on this hardware/loader combination MTP nets out slightly slower than plain decoding — the draft head's acceptance is low and the sparse-MoE verify step can't amortize its weight reads across draft tokens. I ran it against draft depth, the probability threshold, and draft-head precision; none of them turned it into a win here. So I ship it with MTP off. If you want the model's advertised MTP speedup, run it on a CUDA/vLLM stack instead of this one.

License

Derived from deepseek-ai/DeepSeek-V4-Flash-0731; the original model's license applies (see license_link). This upload is only a quantization — all capabilities and limitations are the base model's.

<!-- PEER-TABLE:START -->

Other public builds of this model

Compiled from Hugging Face repository metadata — file sizes, shipped files, quant variant as named by each repo. No third-party build was run or benchmarked here, so this table makes no speed or quality claim about any of them. It is here so you can see the size and format options at a glance and pick what fits your hardware.

Base model: deepseek-ai/DeepSeek-V4-Flash-0731. Generated from Hub metadata; download counts move over time.

<!-- PEER-TABLE:END --> <!-- CREDITS:START -->

Acknowledgements

This build would not exist without the work below. Please star and follow these projects — the quantisation format used here is their engineering, not mine.

ROCmFPX — maintained by charlie12345 / caf The ROCmFP4 / ROCmFPX tensor formats (ggml types 100–106) exist only in this fork. Every ROCmFP4 file in this repository was produced with its llama-quantize, and runs on its runtime. The fork also credits collaborators ciru-ai, Tom Turney, PlunderStruck and Aydan S., and acknowledges AMD for hardware support. Licensed MIT, based on upstream llama.cpp.

llama.cpp — ggml-org and contributors The inference engine, GGUF format and conversion tooling everything here is built on.

AMD ROCm The compute platform these builds target — ROCm 7.2.4 on gfx1151 / Radeon 8060S.

Base model authors — see base_model in the metadata above; all model weights, licences and capabilities are theirs. This repository contributes quantisation and measurement only.

If you use these files, please credit ROCmFPX alongside this repository.

<!-- CREDITS:END -->