Instructions to use maczzzzzz/Ornith-1.0-9b-ROCmFPX-STRIX_LEAN-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use maczzzzzz/Ornith-1.0-9b-ROCmFPX-STRIX_LEAN-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="maczzzzzz/Ornith-1.0-9b-ROCmFPX-STRIX_LEAN-GGUF",
	filename="Ornith-1.0-9b-ROCmFPX-STRIX_LEAN.gguf",
)

llm.create_chat_completion(
	messages = "No input example has been defined for this model task."
)

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use maczzzzzz/Ornith-1.0-9b-ROCmFPX-STRIX_LEAN-GGUF with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf maczzzzzz/Ornith-1.0-9b-ROCmFPX-STRIX_LEAN-GGUF
# Run inference directly in the terminal:
llama cli -hf maczzzzzz/Ornith-1.0-9b-ROCmFPX-STRIX_LEAN-GGUF

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf maczzzzzz/Ornith-1.0-9b-ROCmFPX-STRIX_LEAN-GGUF
# Run inference directly in the terminal:
llama cli -hf maczzzzzz/Ornith-1.0-9b-ROCmFPX-STRIX_LEAN-GGUF

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf maczzzzzz/Ornith-1.0-9b-ROCmFPX-STRIX_LEAN-GGUF
# Run inference directly in the terminal:
./llama-cli -hf maczzzzzz/Ornith-1.0-9b-ROCmFPX-STRIX_LEAN-GGUF

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf maczzzzzz/Ornith-1.0-9b-ROCmFPX-STRIX_LEAN-GGUF
# Run inference directly in the terminal:
./build/bin/llama-cli -hf maczzzzzz/Ornith-1.0-9b-ROCmFPX-STRIX_LEAN-GGUF

Use Docker

docker model run hf.co/maczzzzzz/Ornith-1.0-9b-ROCmFPX-STRIX_LEAN-GGUF

LM Studio
Jan
Ollama
How to use maczzzzzz/Ornith-1.0-9b-ROCmFPX-STRIX_LEAN-GGUF with Ollama:
```
ollama run hf.co/maczzzzzz/Ornith-1.0-9b-ROCmFPX-STRIX_LEAN-GGUF
```

Unsloth Studio

How to use maczzzzzz/Ornith-1.0-9b-ROCmFPX-STRIX_LEAN-GGUF with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for maczzzzzz/Ornith-1.0-9b-ROCmFPX-STRIX_LEAN-GGUF to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for maczzzzzz/Ornith-1.0-9b-ROCmFPX-STRIX_LEAN-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for maczzzzzz/Ornith-1.0-9b-ROCmFPX-STRIX_LEAN-GGUF to start chatting

How to use maczzzzzz/Ornith-1.0-9b-ROCmFPX-STRIX_LEAN-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf maczzzzzz/Ornith-1.0-9b-ROCmFPX-STRIX_LEAN-GGUF

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "maczzzzzz/Ornith-1.0-9b-ROCmFPX-STRIX_LEAN-GGUF"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use maczzzzzz/Ornith-1.0-9b-ROCmFPX-STRIX_LEAN-GGUF with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf maczzzzzz/Ornith-1.0-9b-ROCmFPX-STRIX_LEAN-GGUF

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default maczzzzzz/Ornith-1.0-9b-ROCmFPX-STRIX_LEAN-GGUF

Run Hermes

hermes

Atomic Chat new

OpenClaw new

How to use maczzzzzz/Ornith-1.0-9b-ROCmFPX-STRIX_LEAN-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf maczzzzzz/Ornith-1.0-9b-ROCmFPX-STRIX_LEAN-GGUF

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "maczzzzzz/Ornith-1.0-9b-ROCmFPX-STRIX_LEAN-GGUF" \
  --custom-provider-id llama-cpp \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

Docker Model Runner
How to use maczzzzzz/Ornith-1.0-9b-ROCmFPX-STRIX_LEAN-GGUF with Docker Model Runner:
```
docker model run hf.co/maczzzzzz/Ornith-1.0-9b-ROCmFPX-STRIX_LEAN-GGUF
```

Lemonade

How to use maczzzzzz/Ornith-1.0-9b-ROCmFPX-STRIX_LEAN-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull maczzzzzz/Ornith-1.0-9b-ROCmFPX-STRIX_LEAN-GGUF

Run and chat with the model

lemonade run user.Ornith-1.0-9b-ROCmFPX-STRIX_LEAN-GGUF-{{QUANT_TAG}}

List all available models

lemonade list

Ornith-1.0-9B ROCmFPX STRIX_LEAN — GGUF

ROCmFPX Q4_0_ROCMFP4_STRIX_LEAN quant of deepreinforce-ai/Ornith-1.0-9B (qwen35 hybrid SSM+attention, 8.95 B params, 262 144 native ctx).

Built with charlie12345/ROCmFPX on a Radeon RX 9060 XT 16 GB (gfx1200), ROCm 7.2.3, NixOS 25.11. Quantized 2026-06-27 with build commit 11d76c2.

File	Size	Quant	BPW
`Ornith-1.0-9b-ROCmFPX-STRIX_LEAN.gguf`	4.72 GB	`Q4_0_ROCMFP4_STRIX_LEAN` (4-bit ROCmFP4 + Strix K/V + Q5_K embed)	4.42

This is not a stock llama.cpp quant; you need a ROCmFPX build of llama-server / llama-cli / llama-quantize to load it. The ROCmFP4 weight format is unknown to stock llama.cpp and will fail with unknown quantization.

Scope of these benchmarks — read this first

These numbers are a light baseline, not a thorough ROCmFPX evaluation. The mesh's bench framework is built for production agent workload regression-detection on the local stack, not for the kind of multi-axis sweep that upstream quant maintainers typically publish. Specifically:

Harness scope is bounded. The numbers below come from llama-bench ctx sweeps + the mesh's mesh_eval (6 tests, 4 deterministic + throughput) + hermes_loop_eval (5 agent scenarios). That's a regression suite, not a quality benchmark — it answers "does this quant still serve the mesh's agent stack correctly," not "is this the best possible 4-bit ROCmFP4 quant of this model."
Sample sizes are small. Throughput numbers are 3 reps on a single GPU; correctness is 5 prompts × 16 tokens; agent loop is 5 scenarios with one-shot generation. None of these are powered for statistical significance on a per-token level.
No perplexity / wikitext / MMLU / GSM8K. The mesh's stack isn't a quality benchmark — those are upstream ROCmFPX's territory. If you need a quality signal, charlie's own validation ladder or an lm-eval-harness run is the right tool.
Single GPU class. All measurements are on a 16 GB RDNA4 (RX 9060 XT, gfx1200). No Strix unified-memory, no CDNA, no multi-GPU, no Vulkan, no CUDA. Cross-hardware generalization is not implied.
No human eval. Quality is "byte-identical on factual / short deterministic outputs, divergent on high-entropy creative generation" — which is expected for any 4-bit quant, not a quality verdict on this one specifically.

What this IS good for: a quick signal that the quant (a) loads, (b) runs at sane throughput, (c) doesn't break the mesh's agent tool-calling, (d) scales predictably with context. What this is NOT good for: claiming "this is the best quant of this model," reproducing academic benchmark results, or substituting for upstream's validation work.

If you want the rigorous version, charlie's own ROCmFPX brief + the model's stock GGUF variants (e.g. bartowski/deepreinforce-ai_Ornith-1.0-9B-GGUF) are the place to look.

What we measured

Hardware: Node B, AMD Ryzen 9 5900XT 16-core, Radeon RX 9060 XT 16 GB (gfx1200), ROCm 7.2.3, NixOS 25.11 Software: charlie12345/ROCmFPX main @ 11d76c2 Source GGUF: ornith-1.0-9b-bf16.gguf (BF16, 17.9 GB) from deepreinforce-ai/Ornith-1.0-9B Same-class baseline: stock llama.cpp b9608 Q4_K_M quantized from the same BF16 source

Throughput vs stock Q4_K_M

llama-bench 3 reps, q8_0 KV, fa=on, ngpu-layers=99. Raw JSON: BENCH-strix-lean-ctx-sweep.json and BENCH-stock-Q4_K_M-ctx-sweep.json.

Ctx	STRIX_LEAN pp (t/s)	Q4_K_M pp (t/s)	Δ pp	STRIX_LEAN tg (t/s)	Q4_K_M tg (t/s)	Δ tg
4 K	1903	1607	+18 %	48.0	46.2	+4 %
8 K	1756	1513	+16 %	48.0	46.2	+4 %
16 K	1531	1341	+14 %	48.0	46.2	+4 %
32 K	1215	1093	+11 %	48.1	46.2	+4 %
64 K	862	798	+8 %	48.1	46.2	+4 %

Findings (small-sample, 3 reps — see the scope caveat above):

STRIX_LEAN beats stock Q4_K_M at every ctx tested on prompt processing.
Decode throughput is ~48 t/s across 4 K → 64 K ctx — high-context scaling claim is consistent with the flat-line observation, but 3 reps is too small to claim a tight bound.
Prompt-processing edge narrows as ctx grows (KV cache dominates at 64 K).
The 4 % decode delta is within the noise of 3-rep llama-bench on a 16 GB card; the more interesting signal is the +18 % pp at 4 K and the 10 % smaller file.

KV cache type sweep (added v0.5.134, head_dim=128)

131 K ctx, fa=on, kv-unified, -np 1:

KV type	VRAM	gen t/s
q8_0 (baseline)	8.7 GB	46.3
turbo4 (winner)	7.6 GB	46.6
turbo3	7.4 GB	44.3
q4_0_rocmfp4	7.7 GB	42.6
q4_0_rocmfp4_fast	7.6 GB	41.9

turbo4 is the production default for any head_dim=128 model in the ROCmFPX build. -1.1 GB VRAM, same speed. The turbo3/4 KV types are TheTom's turboquant, absorbed into ROCmFPX main via PlunderStruck commits d859c9e + d0141e8.

Correctness — STRIX_LEAN vs stock Q4_K_M (5 prompts × 16 tokens, top-10 logprobs, sequential A/B)

Prompt	Argmax match	Mean KL	Text
Capital of France	16 / 16 (100 %)	0.35	identical
Fibonacci code	5 / 13 (38 %)	11.6	divergent
Story opener	1 / 16 (6 %)	16.4	divergent
15 × 37 math	16 / 16 (100 %)	0.11	identical
SQL injection	14 / 16 (87 %)	0.21	near-identical
TOTAL	67.5 %	5.51 weighted	—

Byte-identical on factual / short deterministic outputs (KL < 0.4, argmax 87-100 %), high divergence on open-ended creative generation (KL 11-16, argmax 6-38 %). Divergence correlates with prompt entropy — where multiple tokens are near-equal, greedy argmax flips and KL amplifies. This is expected for any 4-bit quant.

Agent / loop validation (raw JSONs included)

mesh_eval.py 4 deterministic tests + throughput (raw-mesh-eval-ornith-strix-lean.json):

Test	Result
`gibberish` (no degenerate repetition)	OK
`thinking_leak` (no `<think>` leakage)	CLEAN
`tool_calling` (single tool call, valid args)	PASS — `get_weather(location=Tokyo)`
`coding` (merge_sorted_lists, runs + passes test)	PASS
`uncensored` (no refusal on security-tools question)	PASS
`throughput` (3×256-token gen, gen t/s mean)	47.1 t/s (±0.1)
`overall_status`	PASS, 4/4

hermes_loop_eval.py 5 scenarios (raw-hermes-loop-ornith-strix-lean.json):

Scenario	Result
`single` (one tool call)	PASS — final answer correct
`chained` (calc → use result)	PASS — `15 × 37 = 555`
`multi_step` (compare 2 cities)	PASS — table + conclusion
`search` (web search + extract)	PASS — Eiffel Tower height
`error_recovery` (file not found)	PARTIAL — model says the file doesn't exist (factually correct) but the test's strict `final_answer_correct: false` flagged it
`overall_status`	PARTIAL, 4/5

The 4/5 loop is the error_recovery scenario's strict-match failure, not a quant defect. The model behaved correctly.

Quick start

# Build llama.cpp with ROCmFPX (the ROCmFPX-fork supports Q4_0_ROCMFP4_STRIX_LEAN weight type)
git clone https://github.com/charlie12345/ROCmFPX
cd ROCmFPX
cmake -S . -B build -DGGML_HIP=ON -DGGML_VULKAN=OFF -DGGML_CUDA=OFF \
  -DCMAKE_HIP_ARCHITECTURES=gfx1200 ...
cmake --build build --target llama-server llama-cli llama-quantize

# Serve (131 072 ctx, turbo4 KV for head_dim=128, fa=on)
./build/bin/llama-server \
  -m Ornith-1.0-9b-ROCmFPX-STRIX_LEAN.gguf \
  -np 1 -c 131072 \
  -ctk turbo4 -ctv turbo4 \
  -kvo -cram 32768 -fa on

Reproduce the quant

# Source (we used the BF16 gguf; any BF16/F16 gguf of the same parent works)
SRC=/mnt/e/llms-models-data/ornith/ornith-1.0-9b-bf16.gguf

# ROCmFPX llama-quantize (preset is built in; see `llama-quantize --help`)
~/ROCmFPX/build-rdna4/bin/llama-quantize \
  $SRC \
  Ornith-1.0-9b-ROCmFPX-STRIX_LEAN.gguf \
  Q4_0_ROCMFP4_STRIX_LEAN

Quantize time: ~6 min cold, <2 min warm-cache. CPU-only, no GPU required.

Files in this repo

File	What it is
`Ornith-1.0-9b-ROCmFPX-STRIX_LEAN.gguf`	The quant. Load only with a ROCmFPX `llama-server`.
`README.md`	This file
`raw-mesh-eval-ornith-strix-lean.json`	`mesh_eval.py` output (2026-06-27 19:51 UTC)
`raw-hermes-loop-ornith-strix-lean.json`	`hermes_loop_eval.py` output (2026-06-27 19:52 UTC)
`BENCH-strix-lean-ctx-sweep.json`	llama-bench ctx sweep (3 reps, 4 K → 64 K)
`BENCH-stock-Q4_K_M-ctx-sweep.json`	Same sweep on the stock baseline
`BENCH-kv-type-sweep.txt`	KV cache type comparison (q8_0, turbo3, turbo4, q4_0_rocmfp4, q4_0_rocmfp4_fast)
`quant-command.sh`	The exact `llama-quantize` invocation used

What's NOT in this repo (caveats)

Stock llama.cpp will not load this file. The ROCmFP4 weight format is unique to charlie12345/ROCmFPX. Use that fork's llama-server/llama-cli/llama-quantize.
No CUDA / non-AMD GPU bench. All measurements are RDNA4 (gfx1200). Vulkan path on RDNA4 has a known upstream regression (charlie12345/rocmfp4-llama issue #6) — we did not test it.
system_fingerprint will be b1-11d76c2 when served by the ROCmFPX build (verified on prior bench runs in this corpus). If you see a different fingerprint, the wrong binary loaded the file.
No multi-GPU / tensor-parallel bench. 9 B params at 4.7 GB fits comfortably on a single 16 GB card; no need to split.
No MTP / speculative-decode bench on this file. Ornith 1.0 9B does not ship with MTP draft heads.
No vision/multimodal test. Ornith 1.0 9B is text-only; the mesh_eval vision test was skipped (HTTP 500 = expected for this model class).

Provenance

Source model: deepreinforce-ai/Ornith-1.0-9B — qwen35 hybrid SSM+attention, 8.95 B params, native ctx 262 144
Source model license: MIT (https://huggingface.co/deepreinforce-ai/Ornith-1.0-9B/blob/main/LICENSE)
Quantizer: charlie12345/ROCmFPX main @ 11d76c2 (2026-06-27)
Quantizer license: MIT
Build hardware: Node B, AMD Ryzen 9 5900XT 16-core, Radeon RX 9060 XT 16 GB (gfx1200), ROCm 7.2.3, NixOS 25.11
Build tooling: NixOS 25.11, ROCm store paths dynamic-discovered. See the meshina repo's references/nixos-rocm-external-build-recipe.md for the build env setup.
Bench harnesses: scripts/mesh-bench/mesh_eval.py + scripts/mesh-bench/hermes_loop_eval.py from the meshina repo
Original bench report: raw/benchmarks/2026-06-27-rocmfpx-validation/briefs/2026-06-27-ornith-rocmfpx-validation.md in the meshina repo

License

The Ornith 1.0 9B parent model is MIT (per its HF model card).
The charlie12345/ROCmFPX quantizer is MIT.
The GGUF in this repo is a derivative of the MIT-licensed parent, produced with the MIT-licensed quantizer. The MIT license is preserved.