Instructions to use kingjones777/PhoneLLM-Alpha-1-ROCmFP4-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use kingjones777/PhoneLLM-Alpha-1-ROCmFP4-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 kingjones777/PhoneLLM-Alpha-1-ROCmFP4-GGUF:Q4_0_ROCMFP
# Run inference directly in the terminal:
llama cli -hf kingjones777/PhoneLLM-Alpha-1-ROCmFP4-GGUF:Q4_0_ROCMFP

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf kingjones777/PhoneLLM-Alpha-1-ROCmFP4-GGUF:Q4_0_ROCMFP
# Run inference directly in the terminal:
llama cli -hf kingjones777/PhoneLLM-Alpha-1-ROCmFP4-GGUF:Q4_0_ROCMFP

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 kingjones777/PhoneLLM-Alpha-1-ROCmFP4-GGUF:Q4_0_ROCMFP
# Run inference directly in the terminal:
./llama-cli -hf kingjones777/PhoneLLM-Alpha-1-ROCmFP4-GGUF:Q4_0_ROCMFP

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 kingjones777/PhoneLLM-Alpha-1-ROCmFP4-GGUF:Q4_0_ROCMFP
# Run inference directly in the terminal:
./build/bin/llama-cli -hf kingjones777/PhoneLLM-Alpha-1-ROCmFP4-GGUF:Q4_0_ROCMFP

Use Docker

docker model run hf.co/kingjones777/PhoneLLM-Alpha-1-ROCmFP4-GGUF:Q4_0_ROCMFP

LM Studio
Jan

vLLM

How to use kingjones777/PhoneLLM-Alpha-1-ROCmFP4-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "kingjones777/PhoneLLM-Alpha-1-ROCmFP4-GGUF"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "kingjones777/PhoneLLM-Alpha-1-ROCmFP4-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/kingjones777/PhoneLLM-Alpha-1-ROCmFP4-GGUF:Q4_0_ROCMFP

Ollama
How to use kingjones777/PhoneLLM-Alpha-1-ROCmFP4-GGUF with Ollama:
```
ollama run hf.co/kingjones777/PhoneLLM-Alpha-1-ROCmFP4-GGUF:Q4_0_ROCMFP
```
Unsloth Desktop

How to use kingjones777/PhoneLLM-Alpha-1-ROCmFP4-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf kingjones777/PhoneLLM-Alpha-1-ROCmFP4-GGUF:Q4_0_ROCMFP

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "kingjones777/PhoneLLM-Alpha-1-ROCmFP4-GGUF:Q4_0_ROCMFP"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use kingjones777/PhoneLLM-Alpha-1-ROCmFP4-GGUF with Docker Model Runner:
```
docker model run hf.co/kingjones777/PhoneLLM-Alpha-1-ROCmFP4-GGUF:Q4_0_ROCMFP
```

Lemonade

How to use kingjones777/PhoneLLM-Alpha-1-ROCmFP4-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull kingjones777/PhoneLLM-Alpha-1-ROCmFP4-GGUF:Q4_0_ROCMFP

Run and chat with the model

lemonade run user.PhoneLLM-Alpha-1-ROCmFP4-GGUF-Q4_0_ROCMFP

List all available models

lemonade list

Hermes Agent

How to use kingjones777/PhoneLLM-Alpha-1-ROCmFP4-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 kingjones777/PhoneLLM-Alpha-1-ROCmFP4-GGUF:Q4_0_ROCMFP

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 kingjones777/PhoneLLM-Alpha-1-ROCmFP4-GGUF:Q4_0_ROCMFP

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use kingjones777/PhoneLLM-Alpha-1-ROCmFP4-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf kingjones777/PhoneLLM-Alpha-1-ROCmFP4-GGUF:Q4_0_ROCMFP

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 "kingjones777/PhoneLLM-Alpha-1-ROCmFP4-GGUF:Q4_0_ROCMFP" \
  --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"

PhoneLLM-Alpha-1-ROCmFP4-GGUF / README.md

kingjones777

add model card

e929d06 verified 3 days ago

preview code

Raw

History Blame Contribute Delete

9.99 kB

	---
	license: bsd-2-clause
	base_model:
	- pipecat-ai/phonellm-alpha-1
	- nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16
	base_model_relation: quantized
	pipeline_tag: text-generation
	library_name: gguf
	language:
	- en
	tags:
	- gguf
	- rocmfp4
	- rocmfpx
	- strix-halo
	- gfx1151
	- vulkan
	- nemotron
	- mixture-of-experts
	- voice-agent
	- tool-use
	- function-calling
	- pipecat
	---

	# PhoneLLM Alpha 1 — ROCmFP4 for AMD Strix Halo (gfx1151)

	> The model here is not our work. PhoneLLM Alpha 1 is by **[Daily](https://www.daily.co/) / the
	> [Pipecat](https://www.pipecat.ai/) team** — [`pipecat-ai/phonellm-alpha-1`](https://huggingface.co/pipecat-ai/phonellm-alpha-1) —
	> a full-parameter fine-tune of [NVIDIA Nemotron 3 Nano 30B-A3B](https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16).
	> This repository adds only the ROCmFP4/ROCmFPX quantisation ladder for AMD Strix Halo and the measurements
	> below. Go star their repo. Pipecat also ship an official NVFP4 build for NVIDIA Blackwell:
	> [`pipecat-ai/phonellm-alpha-1-nvfp4`](https://huggingface.co/pipecat-ai/phonellm-alpha-1-nvfp4).

	**A hybrid Mamba-Transformer MoE voice-agent model — 30B total, 3.5B active — quantised to run on a
	Ryzen AI Max+ 395 (Radeon 8060S, gfx1151). One binary, both backends: HIP (ROCm) and Vulkan are a runtime
	`-dev` flag, not a rebuild.**

	PhoneLLM is built for one job: call the right tool at the right time, with thinking disabled, at
	phone-call latency. That shapes how we verified it — see Verification.

	## Which file should I use?

	\| File \| ftype \| Size \| Head \| Notes \|
	\|---\|---\|---:\|---\|---\|
	\| `Q4_0_ROCMFP4_STRIX_LEAN` \| 106 \| 15.91 GiB \| q8_0 \| Flagship — start here. Smallest tier that keeps full tool behaviour. Also published on its own: [`PhoneLLM-Alpha-1-ROCmFP4-STRIX-LEAN-GGUF`](https://huggingface.co/kingjones777/PhoneLLM-Alpha-1-ROCmFP4-STRIX-LEAN-GGUF) \|
	\| `Q4_0_ROCMFP4_FAST` \| 103 \| 15.83 GiB \| q8_0 \| Smallest file; same tool score as the flagship \|
	\| `Q4_0_ROCMFP4_COHERENT` \| 102 \| 16.91 GiB \| q8_0 \| Highest-precision 4-bit tier \|
	\| `Q6_0_ROCMFPX_AGENT` \| 114 \| 27.26 GiB \| q8_0 \| 6-bit \|
	\| `Q8_0_ROCMFPX` \| 111 \| 30.37 GiB \| q8_0 \| Plain Q8 reference \|
	\| `Q8_0_ROCMFPX_AGENT` \| 115 \| 30.84 GiB \| q8_0 \| 8-bit, agent-tuned tensor map \|

	> Take the 15.91 GiB flagship. Across our probe the 30 GiB Q8 tiers score no better than the 16 GiB
	> 4-bit tiers (see Verification). On a 128 GB Strix Halo that leaves real headroom to co-host your ASR and
	> TTS models on the same box — which is the point, since PhoneLLM is the LLM stage of a voice pipeline, not
	> a speech model (it is text-in / text-out; you still need STT and TTS).

	## Quick start

	```bash
	llama-server -m PhoneLLM-Alpha-1-Q4_0_ROCMFP4_STRIX_LEAN.gguf \
	-dev ROCm0 -fa on -ngl 999 -fit off -np 1 \
	-c 32768 -b 4096 -t 8 --jinja \
	--host 0.0.0.0 --port 8080
	```

	Run it the way Pipecat recommend the source model: `temperature=0` and thinking disabled.

	```json
	{"chat_template_kwargs": {"enable_thinking": false}}
	```

	llama.cpp resolves this model's chat format as `peg-native`; tool calls come back as proper
	`tool_calls` on `/v1/chat/completions` with `--jinja`.

	### ⛔ You need a ROCmFPX build — stock llama.cpp will NOT load these files

	ROCmFP4/ROCmFPX use ggml tensor types 100–119; upstream's table stops at 43. Build with both backends:

	```bash
	cmake -S . -B build-hipvk -DCMAKE_BUILD_TYPE=Release \
	-DGGML_HIP=ON -DGGML_VULKAN=ON -DGGML_NATIVE=ON \
	-DVulkan_GLSLC_EXECUTABLE=/usr/bin/glslc -DAMDGPU_TARGETS=gfx1151
	cmake --build build-hipvk -j
	build-hipvk/bin/llama-server --list-devices # must list ROCm0 AND Vulkan0
	```

	---

	## Head protection — why every tier uses a `q8_0` head

	Our usual ladder protects `output.weight` with `q6_K` on the 4-bit tiers. **That is impossible on this
	model. `hidden_size` is 2688**, and K-quants use 256-element superblocks:

	```
	2688 % 256 = 128 → ggml.c:8024: GGML_ASSERT(start % type_traits[type].blck_size == 0) failed
	[1/401] output.weight - [2688, 131072], bf16, converting to q6_K .. SIGABRT
	```

	`q8_0` uses 32-element blocks and `2688 % 32 == 0`, so every tier here carries a `q8_0` head —
	higher precision than our usual q6_K, at a cost of roughly +88 MB on the head tensor.

	We caught this as a clean natural experiment in a single run: the three `q6_K`-head tiers aborted in ~2 s
	while the `q8_0`-head tier built normally, same source, same binary, same moment. Note that
	`llama-quantize --dry-run` does not catch it — the dry run planned all 401 tensors and printed a clean
	`60247 MiB → 17223 MiB (4.58 BPW)` summary. The assert only fires once real data is written.

	---

	## Verification

	Every tier is checked for load, coherence, and — because this is the whole point of PhoneLLM —
	tool calling, using the vendor-recommended mode (`temperature=0`, `enable_thinking: false`).

	The tool probe is deliberately adversarial: it includes a case where the model must not call anything,
	and a multi-turn case, because the failure Pipecat built this model to avoid is an agent that says
	"yes, I've booked that" without emitting a call.

	All tiers, greedy (`temperature=0`), `enable_thinking: false`, `--jinja`, chat format `peg-native`.

	\| Tier \| Size \| Loads \| Coherent \| Tool probe \|
	\|---\|---:\|---\|---\|---\|
	\| `Q4_0_ROCMFP4_STRIX_LEAN` \| 15.91 GiB \| ✅ ~10 s \| ✅ \| 3/5 \|
	\| `Q4_0_ROCMFP4_FAST` \| 15.83 GiB \| ✅ ~10 s \| ✅ \| 3/5 \|
	\| `Q4_0_ROCMFP4_COHERENT` \| 16.91 GiB \| ✅ ~10 s \| ✅ \| 2/5 \|
	\| `Q6_0_ROCMFPX_AGENT` \| 27.26 GiB \| ✅ ~20 s \| ✅ \| 3/5 \|
	\| `Q8_0_ROCMFPX_AGENT` \| 30.84 GiB \| ✅ ~25 s \| ✅ \| 2/5 \|
	\| `Q8_0_ROCMFPX` \| 30.37 GiB \| ✅ ~20 s \| ✅ \| 3/5 \|
	\| BF16 source (control) \| 58.8 GiB \| ✅ \| ✅ \| 1/5 \|

	6/6 tiers load and stay coherent. There is no precision-dependent degradation: the 30 GiB Q8 tiers score
	the same as the 16 GiB 4-bit tiers, and every quantised tier scores at or above the BF16 control. If
	quantisation were damaging tool calling, the Q8 tiers would lead. They do not — so pick on size.

	Flagship detail (`STRIX_LEAN`), 5 adversarial cases:

	```
	PASS booking book_table {"name":"Chen","party_size":2,"time":"19:00"} ← normalised "7pm" → 19:00
	PASS escalate transfer_to_human {"reason":"Customer has called multiple times ..."}
	PASS no-tool (correctly emitted NO call)
	FAIL availability (no call — the one unambiguous miss)
	FAIL multiturn check_availability {"date":"Saturday","party_size":4}
	```

	Read `3/5` carefully — the rubric is strict and opinionated. The `multiturn` "failure" is the model
	checking availability before booking, which is defensible agent behaviour; we counted it wrong because our
	expected answer demanded a booking. The `no-tool` pass matters most: the model declined to invent a call when
	none was warranted, which is the failure mode PhoneLLM exists to avoid. Treat these as a **smoke test that the
	tool path survives quantisation**, not as a benchmark score — for a real score use Pipecat's PhoneBench.

	### ⚠️ An honest limitation: we could not establish a BF16 baseline on this hardware

	We ran the BF16 GGUF as a control arm and it misbehaves on gfx1151 when tools are attached — the same
	prompt that a quantised tier answers with a correct `book_table` call returns, from BF16, either a
	degenerate repetition loop or an unrelated non-sequitur. Without tools, BF16 is coherent.

	So we can report what the quantised tiers do, but we cannot publish a "delta vs BF16" the way
	Pipecat report NVFP4 (PhoneBench 72.06 → 71.51). Anyone quoting a quality delta for these files against
	BF16 on this hardware would be quoting a broken control. We have not root-caused it (candidates: gfx1151
	bf16 compute, or the `peg-native` tool-template path); it is flagged here rather than papered over.

	---

	## Reproduction block

	A number without its binary is a rumour.

	\| \| \|
	\|---\|---\|
	\| Host \| Ryzen AI Max+ 395 (Strix Halo), Radeon 8060S, gfx1151, 128 GB unified \|
	\| Build \| ROCmFPX fork @ `e7712358806055c70a9753b070202b0cc7c637e3` \|
	\| \| `GGML_HIP=ON GGML_VULKAN=ON GGML_NATIVE=ON`, Release, `AMDGPU_TARGETS=gfx1151` \|
	\| `llama-server` sha256 \| `e860b763d5e8f496ddb4b01236d8a8f37ae09dda456b5d076ab3abe2d9deb9fc` \|
	\| Source \| `pipecat-ai/phonellm-alpha-1`, 13 safetensors shards, 58.8 GiB \|
	\| Converted \| `convert_hf_to_gguf.py --outtype bf16` → 401 tensors, 63.18 GB, arch `nemotron_h_moe` \|
	\| Quantise \| `llama-quantize --output-tensor-type q8_0 <bf16> <out> <ftype> 12` \|
	\| Serve (verification) \| `-dev ROCm0 -fa on -ngl 999 -fit off -np 1 -c 32768 -b 4096 -t 8 --jinja` \|

	---

	## Not measured

	- Perplexity (the source is a voice-agent fine-tune; wikitext PPL is a poor proxy and we would rather
	publish nothing than a misleading number).
	- PhoneBench — that is Pipecat's harness; we did not run it.
	- Vision — text-only model, no projector.
	- Context beyond 32768 (the source supports 262144).
	- Decode throughput per tier.

	## License and attribution

	Released under BSD 2-Clause, matching the source. The source is itself a derivative of an
	NVIDIA Nemotron Open Model License work — see `LICENSE_NVIDIA.txt` in the upstream repo.

	- Model: [`pipecat-ai/phonellm-alpha-1`](https://huggingface.co/pipecat-ai/phonellm-alpha-1) — Daily / Pipecat.
	- Base: [`nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16`](https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16) — NVIDIA.
	- This repository contributes only the ROCmFP4/ROCmFPX quantisation ladder and the measurements above.

	## Acknowledgements

	Daily / Pipecat for PhoneLLM and for publishing an honest PhoneBench methodology. NVIDIA for
	Nemotron 3 Nano and the hybrid Mamba-Transformer architecture. The ROCmFPX project for the FP4/FPX
	tensor types and the Strix Halo kernels that make these files possible.