LFM2-700M

Mid-sized 700M parameter model for deploying on most devices

← Back to Text Models

LFM2-700M is a compact model balancing capability and efficiency. Suitable for deployment on a wide range of devices including phones, tablets, and laptops with limited resources.

HFGGUFMLXONNX

Specifications

Property	Value
Parameters	700M
Context Length	32K tokens
Architecture	LFM2 (Dense)

Edge Deployment

Optimized for resource-constrained devices

Low Latency

Fast inference for real-time applications

Fine-tunable

TRL compatible (SFT, DPO, GRPO)

Quick Start

Transformers

Install:

pip install transformers torch

Download & Run:

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("LiquidAI/LFM2-700M", device_map="auto")
tokenizer = AutoTokenizer.from_pretrained("LiquidAI/LFM2-700M")

input_ids = tokenizer.apply_chat_template(
    [{"role": "user", "content": "What is machine learning?"}],
    add_generation_prompt=True, return_tensors="pt"
).to(model.device)

output = model.generate(input_ids, max_new_tokens=256)
print(tokenizer.decode(output[0], skip_special_tokens=True))

llama.cpp

Install:

brew install llama.cpp
pip install huggingface-hub

Download:

huggingface-cli download LiquidAI/LFM2-700M-GGUF \
  lfm2-700m-q4_k_m.gguf --local-dir .

Run:

llama-cli -m lfm2-700m-q4_k_m.gguf \
  -p "What is machine learning?" -n 256

vLLM

Install:

pip install vllm

Run:

from vllm import LLM, SamplingParams

llm = LLM(model="LiquidAI/LFM2-700M")
output = llm.chat("What is machine learning?", SamplingParams(max_tokens=256))
print(output[0].outputs[0].text)