Instructions to use LiquidAI/LFM2-2.6B-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 LiquidAI/LFM2-2.6B-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 LiquidAI/LFM2-2.6B-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf LiquidAI/LFM2-2.6B-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf LiquidAI/LFM2-2.6B-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf LiquidAI/LFM2-2.6B-GGUF:Q4_K_M

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 LiquidAI/LFM2-2.6B-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama cli -hf LiquidAI/LFM2-2.6B-GGUF:Q4_K_M

Use Docker

docker model run hf.co/LiquidAI/LFM2-2.6B-GGUF:Q4_K_M

Llama new
LM Studio
Jan

vLLM

How to use LiquidAI/LFM2-2.6B-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "LiquidAI/LFM2-2.6B-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": "LiquidAI/LFM2-2.6B-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/LiquidAI/LFM2-2.6B-GGUF:Q4_K_M

Ollama
How to use LiquidAI/LFM2-2.6B-GGUF with Ollama:
```
ollama run hf.co/LiquidAI/LFM2-2.6B-GGUF:Q4_K_M
```
Unsloth Desktop

How to use LiquidAI/LFM2-2.6B-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf LiquidAI/LFM2-2.6B-GGUF:Q4_K_M

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": "LiquidAI/LFM2-2.6B-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use LiquidAI/LFM2-2.6B-GGUF with Docker Model Runner:
```
docker model run hf.co/LiquidAI/LFM2-2.6B-GGUF:Q4_K_M
```

Lemonade

How to use LiquidAI/LFM2-2.6B-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull LiquidAI/LFM2-2.6B-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.LFM2-2.6B-GGUF-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use LiquidAI/LFM2-2.6B-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 LiquidAI/LFM2-2.6B-GGUF:Q4_K_M

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 LiquidAI/LFM2-2.6B-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use LiquidAI/LFM2-2.6B-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf LiquidAI/LFM2-2.6B-GGUF:Q4_K_M

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 "LiquidAI/LFM2-2.6B-GGUF:Q4_K_M" \
  --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"

Running LFM2-2.6B Q4_K_M on an iPhone 15: numbers and tips for structured output

by theyplaygames - opened 1 day ago

Discussion

theyplaygames

1 day ago

Sharing results in case they help others running LFM2 on phones.

Setup: LFM2-2.6B-Q4_K_M.gguf (1.6 GB) with llama.cpp b11887 on an iPhone 15 (A16, 6 GB RAM, no Apple Intelligence), Metal, n_ctx 4096, greedy sampling with a GBNF grammar so the output is always valid JSON.

Timings on the phone:

first model load ~20 s, then it stays in memory
short text → 6-field JSON: 4–6 s
voice-note transcript → task list (nested objects): 7–9 s
OCR'd bill (~1,500 chars) → 7 fields incl. summary and steps: 10–14 s

What worked well: extraction and summaries in 8 languages (EN, DE, FR, ES, JA, KO, ZH, AR), and reading a Japanese bill while answering in English. Compared on our transcripts it was nearly as good as Qwen3-4B at two thirds of the size and about twice the speed.

What needed help from outside the model:

totals on receipts (it sometimes picks a line item) → we take the amount next to "Total/合計" with a rule
weekdays → giving today plus the next 7 dates by name fixed "on Tuesday"
answer language → repeat it after the input, not only in the system prompt
long prompts crash llama.cpp unless n_batch >= prompt length (we use n_batch = n_ctx)

Screen recordings (all offline on the phone): https://theyplay.games/ondevice/#real
GBNF from a Swift schema (MIT): https://github.com/theyplay-games/swift-gbnf
llama.cpp xcframework with an iOS Simulator slice (MIT, unchanged b11887): https://github.com/theyplay-games/llama-xcframework

Thanks for the model, it's the best size/quality trade-off we found for phones.

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment