---
library_name: llama.cpp
license: apache-2.0
base_model: ProCreations/grug-35b-v2
tags:
  - rocmfpx
  - gfx1151
  - strix-halo
  - qwen35moe
  - moe
  - rocm
  - amdgpu
  - ROCmFP4
pipeline_tag: image-text-to-text
language:
  - en
  - multilingual
version: '1.0'
date: 2026-08-11
quantized_by: pugant
---

# grug-35b-v2 — ROCmFP4-STRIX_LEAN (Strix Halo / gfx1151)

> Version 1.0 — 2026-08-11

## TL;DR

`grug-35b-v2` (35B params, 3B active per token, Qwen3.5-VL-MoE family) quantized to **`Q4_0_ROCMFP4_STRIX_LEAN`** (type 106 preset, ~4.29 BPW). Tuned for **AMD Strix Halo (gfx1151 / RDNA 3.5)** with the `charlie12345/ROCmFPX` fork of llama.cpp. Runs the full vision + text multimodal model in ~17.3 GiB.

## ⚠️ Critical warnings — read before downloading

- **Requires `charlie12345/ROCmFPX`** fork of llama.cpp (built via the `kyuz0/amd-strix-halo-toolboxes` container). The type 106 (`Q4_0_ROCMFP4_STRIX_LEAN`) tensor format is **INVALID in stock llama.cpp** — it will refuse to load. See [Usage](#usage) below.
- Profiled for **gfx1151 only** (Strix Halo / Ryzen AI Max+ 395, RDNA 3.5). Not tested on other GPUs.
- FP4 here is **software** on RDNA 3.5 (no FP4 silicon units): the win is **bandwidth / memory**, not raw compute throughput. The Strix Halo ceiling on this MoE is bandwidth-bound, which is exactly where FP4 helps.

## Benchmarks

Tested on Strix Halo (AMD Ryzen AI Max+ 395, 128 GB LPDDR5X). Methodology: `llama-bench -ngl 999 -fa on -p 512 -n 128 -mmap 0`. All benchmarks for this model were run in **ROCm containers** (HIP backend). Since then we also benchmarked ROCmFPX quants on the Vulkan (RADV) build of the fork — see the comparison table in [Qwen3.6-35B-A3B-MTP-Q6_0_ROCMFPX](https://huggingface.co/pugant/Qwen3.6-35B-A3B-MTP-Q6_0_ROCMFPX); those numbers are on a different backend and not directly comparable.

| Model | Quant | Size | tg128 (tok/s) | pp512 (tok/s) |
|---|---|---:|---:|---:|
| grug-35b-v2 | **ROCmFP4-STRIX_LEAN** | **17.31 GiB** | **70.92** | **1418** |
| grug-35b-v2 | Q4_K_M (baseline) | 19.70 GiB | 61.18 | — |
| Qwen3.6-35B-A3B (production ref) | ROCmFP4-STRIX_LEAN | 17.31 GiB | 63 | — |

**Speed-up**: +16% vs Q4_K_M (70.92 vs 61.18 tok/s tg128) at **−12% size** (17.31 vs 19.70 GiB). +12% vs the production Qwen3.6-35B-A3B reference.

### System configuration at bench time

Declared for reproducibility:

- **Bare metal host**: Bosgame BeyondMax Series (`bosgame-m5`), Ubuntu 24.04.4 LTS, kernel 7.0.0-28-generic
- **CPU power profile**: `balanced` (`powerprofilesctl get`) — default, **NOT forced to `performance`**. Representative of an out-of-the-box setup.
- **CPU scaling driver**: `amd-pstate-epp`, scaling_governor `performance` (amd-pstate-epp default), EPP `performance`
- **IOMMU / iGPU power**: auto (no manual tuning)

Note: tok/s above were measured on a non-tuned system (power profile `balanced`). Users who set `powerprofilesctl set performance` may see slightly higher numbers.

## Quantization details

Preset `Q4_0_ROCMFP4_STRIX_LEAN` (GGUF file_type 106, ~4.29 bits/weight):

- **Attention K/V** (`blk.*.attn_qkv.weight`, `blk.*.attn_v.weight`) → `q4_0_rocmfp4` (high-precision path for attention state)
- **Token embeddings** (`token_embd.weight`) → `Q5_K` (preserve vocab fidelity)
- **Expert FFN** (`blk.*.ffn_*_exps.weight`) → `q4_0_rocmfp4_fast` (max speed path; the bulk of MoE weights)
- Other tensors → F32 / Q4_0_ROCMFP4_FAST as appropriate

Reference fork: [`charlie12345/ROCmFPX`](https://github.com/charlie12345/ROCmFPX) commit `00d5452`.

## imatrix methodology

Generated with `llama-imatrix` (256 chunks, 16 threads, CPU-only). Calibration text from [`ProCreations/grug-think-v3-10k`](https://huggingface.co/datasets/ProCreations/grug-think-v3-10k) — **public Apache-2.0 dataset, not gated**: anyone can download it to replicate. Many thanks to the grug team for publishing both the model and a clean calibration set.

- 510 entries over 733 tensors
- Warning `partial data 99.61%` during quantization = 1/256 expert not activated in calibration (normal for MoE — see `tools/imatrix/imatrix.cpp` in llama.cpp). Negligible impact.

## Files

| File | Size | Description |
|---|---:|---|
| `grug-35b-v2-ROCmFP4-STRIX_LEAN.gguf` | ~17.32 GiB | Main model (type 106) |
| `mmproj-grug-35b-v2-f16.gguf` | ~857 MB | Vision projector (F16) |
| `imatrix-grug-35b-v2.gguf` | ~183 MB | Importance matrix (for re-quantization) |

## Usage

```bash
# Requires the kyuz0 Strix Halo toolbox (which builds charlie12345/ROCmFPX)
docker run --rm -p 1234:1234 --device /dev/kfd --device /dev/dri \
  -v /path/to/models:/models rocmfpx-llm-service \
  llama-server \
    -m /models/grug-35b-v2-ROCmFP4-STRIX_LEAN.gguf \
    --mmproj /models/mmproj-grug-35b-v2-f16.gguf \
    -ngl 999 -fa on --jinja -c 32768 --host 0.0.0.0 --port 1234
```

Notes:
- **MTP not enabled** for grug. The `mtp_num_hidden_layers` field is 0 in this model (MTP was removed during fine-tuning), so it cannot be activated.
- The `--mmproj` flag is required for the vision tower (multimodal). Without it, text-only still works.

## How to replicate

Textual pipeline only (no published scripts):

1. Build the `docker-llm-service-convert` image from `kyuz0/amd-strix-halo-toolboxes` + `charlie12345/ROCmFPX` (commit `00d5452` or later main HEAD — must contain `MODEL_ARCH.QWEN35MOE`).
2. Download the BF16 safetensors from `ProCreations/grug-35b-v2`.
3. Convert to GGUF with `convert_hf_to_gguf.py` (inside the container).
4. Generate the imatrix with `llama-imatrix` using `ProCreations/grug-think-v3-10k` (256 chunks).
5. Quantize: `llama-quantize <bf16>.gguf <out>.gguf Q4_0_ROCMFP4_STRIX_LEAN 16`.

## Attribution & model tree

```
Qwen3.5-VL-MoE (base architecture)
    └── ornith-ai/Ornith-1.0-35B (MIT)
            └── ProCreations/grug-35b-v2 (Apache-2.0)
                    └── this GGUF (ROCmFP4-STRIX_LEAN)
```

- Base model: [`ProCreations/grug-35b-v2`](https://huggingface.co/ProCreations/grug-35b-v2) (Apache-2.0)
- Quantization fork: [`charlie12345/ROCmFPX`](https://github.com/charlie12345/ROCmFPX) (MIT)
- Container runtime: [`kyuz0/amd-strix-halo-toolboxes`](https://github.com/kyuz0/amd-strix-halo-toolboxes)

## License

Apache-2.0 (inherited from `ProCreations/grug-35b-v2`). Derivative work: original model and its license are preserved. See [`LICENSE`](./LICENSE) and [`NOTICE`](./NOTICE).

## Acknowledgements

Built on the shoulders of giants:

- [kyuz0/amd-strix-halo-toolboxes](https://github.com/kyuz0/amd-strix-halo-toolboxes) — Strix Halo container runtime
- [charlie12345/ROCmFPX](https://github.com/charlie12345/ROCmFPX) — llama.cpp fork with ROCmFP4 presets (type 106)
- [ProCreations](https://huggingface.co/ProCreations) — grug-35b-v2 + the grug-think-v3-10k calibration dataset
- [ornith-ai / DeepReinforce Team](https://huggingface.co/deepreinforce-ai) — Ornith-1.0-35B, the base of grug
- [llama.cpp](https://github.com/ggerganov/llama.cpp) community + Kawrakow (imatrix methodology)
- Hardware: Bosgame BeyondMax Series (Strix Halo bare metal host)

## Limitations & community feedback

- **Speed benchmark only.** No perplexity / MMLU / quality eval is included in this release. The MoE structure is preserved bit-for-bit from the BF16 source except for the quantized tensor formats above; quality is expected to track standard Q4_K_M-class with the ROCmFP4 attention/K-V choices, but this is not measured here.
- Profiled for **gfx1151 only**. Not tested on other GPUs (no Navi 3 / Navi 4 / data-center MI series numbers — feel free to share yours).
- MTP not activated (plain inference).

**We invite the community — especially fellow Strix Halo owners — to test and share quality results. Open a Discussion on this repo.**

## Citation

```bibtex
@misc{grug35b2026,
  title  = {grug-35b-v2},
  author = {ProCreations},
  year   = {2026},
  url    = {https://huggingface.co/ProCreations/grug-35b-v2}
}
```

## Disclaimer

No affiliation with AMD, Qwen, ProCreations, DeepReinforce, unsloth, kyuz0, or charlie12345. Provided as-is, without warranty. Users must comply with the base model license (Apache-2.0).