Instructions to use LiquidAI/LFM2-1.2B-Tool-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use LiquidAI/LFM2-1.2B-Tool-GGUF with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="LiquidAI/LFM2-1.2B-Tool-GGUF")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("LiquidAI/LFM2-1.2B-Tool-GGUF", dtype="auto")

llama-cpp-python

How to use LiquidAI/LFM2-1.2B-Tool-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="LiquidAI/LFM2-1.2B-Tool-GGUF",
	filename="LFM2-1.2B-Tool-F16.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": "What is the capital of France?"
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use LiquidAI/LFM2-1.2B-Tool-GGUF with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf LiquidAI/LFM2-1.2B-Tool-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf LiquidAI/LFM2-1.2B-Tool-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf LiquidAI/LFM2-1.2B-Tool-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf LiquidAI/LFM2-1.2B-Tool-GGUF:Q4_K_M

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 LiquidAI/LFM2-1.2B-Tool-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf LiquidAI/LFM2-1.2B-Tool-GGUF:Q4_K_M

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 LiquidAI/LFM2-1.2B-Tool-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf LiquidAI/LFM2-1.2B-Tool-GGUF:Q4_K_M

Use Docker

docker model run hf.co/LiquidAI/LFM2-1.2B-Tool-GGUF:Q4_K_M

LM Studio
Jan

vLLM

How to use LiquidAI/LFM2-1.2B-Tool-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "LiquidAI/LFM2-1.2B-Tool-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/LFM2-1.2B-Tool-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/LiquidAI/LFM2-1.2B-Tool-GGUF:Q4_K_M

SGLang

How to use LiquidAI/LFM2-1.2B-Tool-GGUF with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "LiquidAI/LFM2-1.2B-Tool-GGUF" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "LiquidAI/LFM2-1.2B-Tool-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "LiquidAI/LFM2-1.2B-Tool-GGUF" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "LiquidAI/LFM2-1.2B-Tool-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use LiquidAI/LFM2-1.2B-Tool-GGUF with Ollama:
```
ollama run hf.co/LiquidAI/LFM2-1.2B-Tool-GGUF:Q4_K_M
```

Unsloth Studio

How to use LiquidAI/LFM2-1.2B-Tool-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 LiquidAI/LFM2-1.2B-Tool-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 LiquidAI/LFM2-1.2B-Tool-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for LiquidAI/LFM2-1.2B-Tool-GGUF to start chatting

How to use LiquidAI/LFM2-1.2B-Tool-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf LiquidAI/LFM2-1.2B-Tool-GGUF:Q4_K_M

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": "LiquidAI/LFM2-1.2B-Tool-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use LiquidAI/LFM2-1.2B-Tool-GGUF with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf LiquidAI/LFM2-1.2B-Tool-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/LFM2-1.2B-Tool-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat new
Docker Model Runner
How to use LiquidAI/LFM2-1.2B-Tool-GGUF with Docker Model Runner:
```
docker model run hf.co/LiquidAI/LFM2-1.2B-Tool-GGUF:Q4_K_M
```

Lemonade

How to use LiquidAI/LFM2-1.2B-Tool-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull LiquidAI/LFM2-1.2B-Tool-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.LFM2-1.2B-Tool-GGUF-Q4_K_M

List all available models

lemonade list

LFM2-1.2B-Tool-Q4_K_M-cuda

by fullstack - opened Nov 19, 2025

Discussion

fullstack

Nov 19, 2025

v1
Model: LFM2-1.2B-Tool-Q4_K_M-cuda

============================================================
AXIS 1: Single Tool - Parameter Scaling

--- Testing 1 parameters ---
Query: Call the test tool with these values: value1...
Calls: ['test_tool(param_1="value1")']
Final: <|tool_call_start|>[test_tool(param_1="value1")]<|tool_call_...
Result: ✓ - Included 1/1 params

--- Testing 2 parameters ---
Query: Call the test tool with these values: value1, value2...
Calls: ['test_tool(param_1="value1", param_2="value2")']
Final: <|tool_call_start|>[test_tool(param_1="mocked_result", param...
Result: ✓ - Included 2/2 params

--- Testing 5 parameters ---
Query: Call the test tool with these values: value1, value2, value3...
Calls: ['test_tool(param_1="value1", param_2="value2", param_3="value3", param_4="value4", param_5="value5")']
Final: <|tool_call_start|>[test_tool(param_1="value1", param_2="val...
Result: ✓ - Included 5/5 params

--- Testing 10 parameters ---
Query: Call the test tool with these values: value1, value2, value3...
Calls: ['test_tool(param_1="value1", param_2="value2", param_3="value3", param_4="value4", param_5="value5", param_6="value6", param_7="value7", param_8="value8", param_9="value9", param_10="value10")']
Final: The provided function test_tool successfully executed with...
Result: ✓ - Included 10/10 params

--- Testing 15 parameters ---
Query: Call the test tool with these values: value1, value2, value3...
Calls: ['test_tool(param_1="value1", param_2="value2", param_3="value3", param_4="value4", param_5="value5", param_6="value6", param_7="value7", param_8="value8", param_9="value9", param_10="value10", param_11="value11", param_12="value12", param_13="value13", param_14="value14", param_15="value15")']
Final: The tool test was successful with the provided parameters. H...
Result: ✓ - Included 15/15 params

--- Testing 20 parameters ---
Query: Call the test tool with these values: value1, value2, value3...
Calls: ['test_tool(param_1="value1", param_2="value2", param_3="value3", param_4="value4", param_5="value5", param_6="value6", param_7="value7", param_8="value8", param_9="value9", param_10="value10", param_11="value11", param_12="value12", param_13="value13", param_14="value14", param_15="value15", param_16="value16", param_17="value17", param_18="value18", param_19="value19", param_20="value20")']
Final: The tool has been tested with the specified parameters. Here...
Result: ✓ - Included 20/20 params

--- Testing 25 parameters ---
Query: Call the test tool with these values: value1, value2, value3...
Calls: ['test_tool(param_1="value1", param_2="value2", param_3="value3", param_4="value4", param_5="value5", param_6="value6", param_7="value7", param_8="value8", param_9="value9", param_10="value10", param_11="value11", param_12="value12", param_13="value13", param_14="value14", param_15="value15", param_16="value16", param_17="value17", param_18="value18", param_19="value19", param_20="value20", param_21="value21", param_22="value22", param_23="value23", param_24="value24", param_25="value25")']
Final: The tool has been tested successfully with the provided para...
Result: ✓ - Included 25/25 params

--- Testing 30 parameters ---
Query: Call the test tool with these values: value1, value2, value3...
Calls: ['test_tool(param_1="value1", param_2="value2", param_3="value3", param_4="value4", param_5="value5", param_6="value6", param_7="value7", param_8="value8", param_9="value9", param_10="value10", param_11="value11", param_12="value12", param_13="value13", param_14="value14", param_15="value15", param_16="value16", param_17="value17", param_18="value18", param_19="value19", param_20="value20", param_21="value21", param_22="value22", param_23="value23", param_24="value24", param_25="value25", param_26="value26", param_27="value27", param_28="value28", param_29="value29", param_30="value30")']
Final: It seems like you're trying to test a tool with a set of par...
Result: ✓ - Included 30/30 params

AXIS 1 SUMMARY:
1 params: ✓ (1/1 included)
2 params: ✓ (2/2 included)
5 params: ✓ (5/5 included)
10 params: ✓ (10/10 included)
15 params: ✓ (15/15 included)
20 params: ✓ (20/20 included)
25 params: ✓ (25/25 included)
30 params: ✓ (30/30 included)

Limit: 30 parameters

============================================================
AXIS 2: Multiple Tools - Simple Parameters

--- Testing 1 tools ---
Query: Use tool_1 with input 'test'...
Calls: ['tool_1(input="test")']
Final: The tool_1 function has been called with the input 'test'. T...
Result: ✓

--- Testing 5 tools ---
Query: Use tool_1 with input 'test'...
Calls: ['tool_1(input="test")']
Final: <|tool_call_start|>[tool_1(input="test")]<|tool_call_end|>...
Result: ✓

--- Testing 10 tools ---
Query: Use tool_1 with input 'test'...
Calls: ['tool_1(input="test")']
Final: <|tool_call_start|>[tool_1(input="test")]<|tool_call_end|>...
Result: ✓

--- Testing 20 tools ---
Query: Use tool_1 with input 'test'...
Calls: ['tool_1(input="test")']
Final: <|tool_call_start|>[tool_1(input="test")]<|tool_call_end|>...
Result: ✓

--- Testing 30 tools ---
Query: Use tool_1 with input 'test'...
Calls: ['tool_1(input="test")']
Final: <|tool_call_start|>[tool_1(input="test")]<|tool_call_end|>...
Result: ✓

--- Testing 50 tools ---
Query: Use tool_1 with input 'test'...
Calls: ['tool_1(input="test")']
Final: <|tool_call_start|>[tool_1(input="test")]<|tool_call_end|>...
Result: ✓

--- Testing 75 tools ---
Query: Use tool_1 with input 'test'...
Calls: ['tool_1(input="test")']
Final: <|tool_call_start|>[tool_1(input="test")]<|tool_call_end|>...
Result: ✓

--- Testing 100 tools ---
Query: Use tool_1 with input 'test'...
Calls: ['tool_1(input="test")']
Final: <|tool_call_start|>[tool_2(input="mocked_result")]<|tool_cal...
Result: ✓

AXIS 2 SUMMARY:
1 tools: ✓
5 tools: ✓
10 tools: ✓
20 tools: ✓
30 tools: ✓
50 tools: ✓
75 tools: ✓
100 tools: ✓

Limit: 100 tools

============================================================
AXIS 3: Sequential Calls - Simple Tools

--- Testing 1 sequential calls ---
Query: Call these tools in order: tool_1...
Calls: ['tool_1()']
Final: <|tool_call_start|>[tool_1(input="mocked_result")]<|tool_cal...
Result: ✓ - Got 1/1 calls, 1 correct

--- Testing 2 sequential calls ---
Query: Call these tools in order: tool_1, tool_2...
Calls: ['tool_1(input="tool_1")', 'tool_2(input="tool_2")']
Final: <|tool_call_start|>[tool_3(input="mocked_result")]<|tool_cal...
Result: ✓ - Got 2/2 calls, 2 correct

--- Testing 3 sequential calls ---
Query: Call these tools in order: tool_1, tool_2, tool_3...
Calls: ['tool_1()', 'tool_2()', 'tool_3()']
Final: <|tool_call_start|>[tool_1(input="mocked_result")]<|tool_cal...
Result: ✓ - Got 3/3 calls, 3 correct

--- Testing 5 sequential calls ---
Query: Call these tools in order: tool_1, tool_2, tool_3, tool_4, t...
Calls: ['tool_1(input="tool_1")', 'tool_2(input="tool_2")', 'tool_3(input="tool_3")', 'tool_4(input="tool_4")', 'tool_5(input="tool_5")']
Final: <|tool_call_start|>[tool_1(input="mocked_result")]<|tool_cal...
Result: ✓ - Got 5/5 calls, 5 correct

--- Testing 7 sequential calls ---
Query: Call these tools in order: tool_1, tool_2, tool_3, tool_4, t...
Calls: ['tool_1(input="tool_1")', 'tool_2(input="tool_2")', 'tool_3(input="tool_3")', 'tool_4(input="tool_4")', 'tool_5(input="tool_5")', 'tool_6(input="tool_6")', 'tool_7(input="tool_7")']
Final: {
"tool_response": {
"content": "{"status": "succes...
Result: ✓ - Got 7/7 calls, 7 correct

--- Testing 10 sequential calls ---
Query: Call these tools in order: tool_1, tool_2, tool_3, tool_4, t...
Calls: ['tool_1()', 'tool_2()', 'tool_3()', 'tool_4()', 'tool_5()', 'tool_6()', 'tool_7()', 'tool_8()', 'tool_9()', 'tool_10()']
Final: {
"tool_response": {
"content": "{"status": "succes...
Result: ✓ - Got 10/10 calls, 10 correct

--- Testing 15 sequential calls ---
Query: Call these tools in order: tool_1, tool_2, tool_3, tool_4, t...
Calls: ['tool_1()', 'tool_2()', 'tool_3()', 'tool_4()', 'tool_5()', 'tool_6()', 'tool_7()', 'tool_8()', 'tool_9()', 'tool_10()', 'tool_11()', 'tool_12()', 'tool_13()', 'tool_14()', 'tool_15()']
Final: {
"tool_response": {
"content": "{"status": "succes...
Result: ✓ - Got 15/15 calls, 15 correct

--- Testing 20 sequential calls ---
Query: Call these tools in order: tool_1, tool_2, tool_3, tool_4, t...
Calls: ['tool_1()', 'tool_2()', 'tool_3()', 'tool_4()', 'tool_5()', 'tool_6()', 'tool_7()', 'tool_8()', 'tool_9()', 'tool_10()', 'tool_11()', 'tool_12()', 'tool_13()', 'tool_14()', 'tool_15()', 'tool_16()', 'tool_17()', 'tool_18()', 'tool_19()', 'tool_20()']
Final: {
"tool_response": {
"content": "{"status": "succes...
Result: ✓ - Got 20/20 calls, 20 correct

AXIS 3 SUMMARY:
1 calls: ✓ (got 1, 1 correct)
2 calls: ✓ (got 2, 2 correct)
3 calls: ✓ (got 3, 3 correct)
5 calls: ✓ (got 5, 5 correct)
7 calls: ✓ (got 7, 7 correct)
10 calls: ✓ (got 10, 10 correct)
15 calls: ✓ (got 15, 15 correct)
20 calls: ✓ (got 20, 20 correct)

Limit: 20 sequential calls

============================================================
FINAL SUMMARY - MODEL LIMITS

Axis 1 (Parameters): 30 max parameters
Axis 2 (Multiple Tools): 100 max tools
Axis 3 (Sequential): 20 max sequential calls

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

LFM2-1.2B-Tool-Q4_K_M-cuda

============================================================TOOL CALLING LIMITS - THREE-AXIS TESTINGServer: http://localhost:8080/v1Model: LFM2-1.2B-Tool-Q4_K_M-cuda

============================================================AXIS 1: Single Tool - Parameter Scaling

Limit: 30 parameters

============================================================AXIS 2: Multiple Tools - Simple Parameters