Instructions to use rcmorano/Qwen3.8-27B-ROCMFPX with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use rcmorano/Qwen3.8-27B-ROCMFPX 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 rcmorano/Qwen3.8-27B-ROCMFPX:Q2_0_ROCMFPX
# Run inference directly in the terminal:
llama cli -hf rcmorano/Qwen3.8-27B-ROCMFPX:Q2_0_ROCMFPX

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf rcmorano/Qwen3.8-27B-ROCMFPX:Q2_0_ROCMFPX
# Run inference directly in the terminal:
llama cli -hf rcmorano/Qwen3.8-27B-ROCMFPX:Q2_0_ROCMFPX

Build from source code

git clone https://github.com/ggml-org/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-app
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama serve -hf rcmorano/Qwen3.8-27B-ROCMFPX:Q2_0_ROCMFPX
# Run inference directly in the terminal:
./build/bin/llama cli -hf rcmorano/Qwen3.8-27B-ROCMFPX:Q2_0_ROCMFPX

Use Docker

docker model run hf.co/rcmorano/Qwen3.8-27B-ROCMFPX:Q2_0_ROCMFPX

Llama new
LM Studio
Jan

Ollama

How to use rcmorano/Qwen3.8-27B-ROCMFPX with Ollama:

ollama run hf.co/rcmorano/Qwen3.8-27B-ROCMFPX:Q2_0_ROCMFPX

Unsloth Desktop

How to use rcmorano/Qwen3.8-27B-ROCMFPX with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf rcmorano/Qwen3.8-27B-ROCMFPX:Q2_0_ROCMFPX

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": "rcmorano/Qwen3.8-27B-ROCMFPX:Q2_0_ROCMFPX"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use rcmorano/Qwen3.8-27B-ROCMFPX with Docker Model Runner:
```
docker model run hf.co/rcmorano/Qwen3.8-27B-ROCMFPX:Q2_0_ROCMFPX
```

Lemonade

How to use rcmorano/Qwen3.8-27B-ROCMFPX with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull rcmorano/Qwen3.8-27B-ROCMFPX:Q2_0_ROCMFPX

Run and chat with the model

lemonade run user.Qwen3.8-27B-ROCMFPX-Q2_0_ROCMFPX

List all available models

lemonade list

Hermes Agent

How to use rcmorano/Qwen3.8-27B-ROCMFPX with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf rcmorano/Qwen3.8-27B-ROCMFPX:Q2_0_ROCMFPX

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 rcmorano/Qwen3.8-27B-ROCMFPX:Q2_0_ROCMFPX

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use rcmorano/Qwen3.8-27B-ROCMFPX with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf rcmorano/Qwen3.8-27B-ROCMFPX:Q2_0_ROCMFPX

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 "rcmorano/Qwen3.8-27B-ROCMFPX:Q2_0_ROCMFPX" \
  --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"

Qwen3.8-27B-ROCMFPX / README.md

rcmorano

Update README.md

11b8dc8 verified 3 months ago

preview code

Raw History Blame Contribute Delete

2.29 kB

---
base_model:
- unsloth/Qwen3.8-27B-GGUF
---
Just plain `llama-quantize`d version of Unsloth's GGUFs :)

My `llama-swap` conf for `Q4_0_ROCMFP4_COHERENT`:

```
    "qwen38":
        name: "qwen38"
        env:
            - "MODEL_ID=qwen38"
        filters:
            stripParams: "temperature, top_k, top_p, repeat_penalty, min_p, presence_penalty"
            setParamsByID:
                "${MODEL_ID}:low-reasoning":
                    chat_template_kwargs:
                        reasoning_effort: low
                "${MODEL_ID}:med-reasoning":
                    chat_template_kwargs:
                        reasoning_effort: medium
                "${MODEL_ID}:instruct-nothink":
                    temperature: 0.7
                    top_p: 0.8
                    min_p: 0.0
                    presence_penalty: 1.5
                    repeat_penalty: 1.0
                    chat_template_kwargs:
                        enable_thinking: false
                        preserve_thinking: false
                "${MODEL_ID}:reasoning-nothink":
                    temperature: 0.85
                    chat_template_kwargs:
                        enable_thinking: false
                        preserve_thinking: false
        proxy: http://host.docker.internal:${PORT}
        cmdStop: docker stop rocmfpx-docker-rocm-llama-${PORT}
        cmd: |
             ${rocmfpx-docker-rocm-llama}
             --model /mnt/models/Qwen3.8-27B-ROCMFPX/Qwen3.8-27B-Q4_0_ROCMFP4_COHERENT.gguf
             --ctx-size 262144
             --ctx-checkpoints 32
             --checkpoint-every-n-tokens 8192
             --seed 69420
             --temperature 1.0
             --top-p 0.95
             --top-k 20
             --min-p 0.0
             --presence_penalty 0.0
             --repeat_penalty 1.0
             --spec-draft-type-k q4_0
             --spec-draft-type-v q4_0
             --cache-type-k f16
             --cache-type-v f16
             --spec-type draft-mtp
             --spec-draft-ngl all
             --spec-draft-n-max 5
             --spec-draft-threads 16
             --spec-draft-threads-batch 32
             --spec-draft-poll 1
             --spec-draft-poll-batch 1
             --chat-template-kwargs '{"preserve_thinking": true}'
```

	---
	base_model:
	- unsloth/Qwen3.8-27B-GGUF
	---

	Just plain `llama-quantize`d version of Unsloth's GGUFs :)

	My `llama-swap` conf for `Q4_0_ROCMFP4_COHERENT`:

	```
	"qwen38":
	name: "qwen38"
	env:
	- "MODEL_ID=qwen38"
	filters:
	stripParams: "temperature, top_k, top_p, repeat_penalty, min_p, presence_penalty"
	setParamsByID:
	"${MODEL_ID}:low-reasoning":
	chat_template_kwargs:
	reasoning_effort: low
	"${MODEL_ID}:med-reasoning":
	chat_template_kwargs:
	reasoning_effort: medium
	"${MODEL_ID}:instruct-nothink":
	temperature: 0.7
	top_p: 0.8
	min_p: 0.0
	presence_penalty: 1.5
	repeat_penalty: 1.0
	chat_template_kwargs:
	enable_thinking: false
	preserve_thinking: false

	"${MODEL_ID}:reasoning-nothink":
	temperature: 0.85
	chat_template_kwargs:
	enable_thinking: false
	preserve_thinking: false
	proxy: http://host.docker.internal:${PORT}
	cmdStop: docker stop rocmfpx-docker-rocm-llama-${PORT}
	cmd: \|
	${rocmfpx-docker-rocm-llama}
	--model /mnt/models/Qwen3.8-27B-ROCMFPX/Qwen3.8-27B-Q4_0_ROCMFP4_COHERENT.gguf
	--ctx-size 262144
	--ctx-checkpoints 32
	--checkpoint-every-n-tokens 8192
	--seed 69420
	--temperature 1.0
	--top-p 0.95
	--top-k 20
	--min-p 0.0
	--presence_penalty 0.0
	--repeat_penalty 1.0
	--spec-draft-type-k q4_0
	--spec-draft-type-v q4_0
	--cache-type-k f16
	--cache-type-v f16
	--spec-type draft-mtp
	--spec-draft-ngl all
	--spec-draft-n-max 5
	--spec-draft-threads 16
	--spec-draft-threads-batch 32
	--spec-draft-poll 1
	--spec-draft-poll-batch 1
	--chat-template-kwargs '{"preserve_thinking": true}'
	```