Instructions to use h2loop-ai/qwen3-0.6b-hexagon with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use h2loop-ai/qwen3-0.6b-hexagon 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 h2loop-ai/qwen3-0.6b-hexagon
# Run inference directly in the terminal:
llama cli -hf h2loop-ai/qwen3-0.6b-hexagon

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf h2loop-ai/qwen3-0.6b-hexagon
# Run inference directly in the terminal:
llama cli -hf h2loop-ai/qwen3-0.6b-hexagon

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 h2loop-ai/qwen3-0.6b-hexagon
# Run inference directly in the terminal:
./llama-cli -hf h2loop-ai/qwen3-0.6b-hexagon

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 h2loop-ai/qwen3-0.6b-hexagon
# Run inference directly in the terminal:
./build/bin/llama-cli -hf h2loop-ai/qwen3-0.6b-hexagon

Use Docker

docker model run hf.co/h2loop-ai/qwen3-0.6b-hexagon

LM Studio
Jan

vLLM

How to use h2loop-ai/qwen3-0.6b-hexagon with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "h2loop-ai/qwen3-0.6b-hexagon"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "h2loop-ai/qwen3-0.6b-hexagon",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/h2loop-ai/qwen3-0.6b-hexagon

Ollama
How to use h2loop-ai/qwen3-0.6b-hexagon with Ollama:
```
ollama run hf.co/h2loop-ai/qwen3-0.6b-hexagon
```
Unsloth Desktop

How to use h2loop-ai/qwen3-0.6b-hexagon with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf h2loop-ai/qwen3-0.6b-hexagon

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": "h2loop-ai/qwen3-0.6b-hexagon"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use h2loop-ai/qwen3-0.6b-hexagon with Docker Model Runner:
```
docker model run hf.co/h2loop-ai/qwen3-0.6b-hexagon
```

Lemonade

How to use h2loop-ai/qwen3-0.6b-hexagon with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull h2loop-ai/qwen3-0.6b-hexagon

Run and chat with the model

lemonade run user.qwen3-0.6b-hexagon-{{QUANT_TAG}}

List all available models

lemonade list

Hermes Agent

How to use h2loop-ai/qwen3-0.6b-hexagon with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf h2loop-ai/qwen3-0.6b-hexagon

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 h2loop-ai/qwen3-0.6b-hexagon

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use h2loop-ai/qwen3-0.6b-hexagon with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf h2loop-ai/qwen3-0.6b-hexagon

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 "h2loop-ai/qwen3-0.6b-hexagon" \
  --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"

qwen3-0.6b-hexagon / results_v81_profile.json

zi-roy

Qwen3-0.6B for Hexagon NPU: A16W4 (llama.cpp) and A16W8 + 16-bit KV (v79, v81)

d725fe3 27 days ago

Raw

History Blame Contribute Delete

725 Bytes

	{
	"target": "Hexagon v81 (SM8850, Snapdragon 8 Elite Gen 5)",
	"method": "AI Hub per-graph profile on real v81 silicon (no Gen 5 device on QDC for end-to-end runs)",
	"qairt": "2.45.0.260326154327",
	"graphs": {
	"part2_token_ar1_cl1024": {"latency_ms": 15.326, "peak_mem_bytes": [77946880, 88260320]},
	"part2_prompt_ar128_cl1024":{"latency_ms": 22.531, "peak_mem_bytes": [64045056, 74921952]},
	"part1_token_ar1_cl1024": {"latency_ms": 0.040, "peak_mem_bytes": [20480, 10073728]}
	},
	"derived": {"decode_tok_s": 65.1, "prefill_tok_s": 5681},
	"caveat": "per-graph latency, not end-to-end. On v79 the same method yields 65.8 tok/s against 72.0 measured end-to-end, so it understates by roughly 9%."
	}