---

license: other

license_name: lfm1.0

license_link: https://huggingface.co/LiquidAI/LFM2.5-350M/blob/main/LICENSE

base_model: hauser458original/lfm2.5-350m-python-math

tags:

- lfm2

- lfm2.5

- liquid

- python

- math

- gguf

- llama.cpp

language:

- en

pipeline_tag: text-generation

---

LFM2.5-350M-Python-Math-GGUF

GGUF quantized versions of hauser458original/lfm2.5-350m-python-math, a Python/math-focused fine-tune of LiquidAI/LFM2.5-350M (instruct) with balanced general chat retention. See the base fine-tune's model card for full training details, evaluation notes, and known limitations.

For use with llama.cpp, Ollama, LM Studio, or any other GGUF-compatible runtime.

Files

| File | Quantization | Approx. size | Notes |

|---|---|---|---|

| lfm2.5-350m-python-math-F16.gguf | F16 | ~700 MB | Full precision, largest, highest fidelity |

| lfm2.5-350m-python-math-Q8_0.gguf | Q8_0 | ~375 MB | Near-lossless, good default if size isn't a concern |

| lfm2.5-350m-python-math-Q5_K_M.gguf | Q5_K_M | ~250 MB | Good balance of size/quality |

| lfm2.5-350m-python-math-Q5_K_S.gguf | Q5_K_S | ~235 MB | Slightly smaller than Q5_K_M, marginal quality trade-off |

| lfm2.5-350m-python-math-Q4_K_M.gguf | Q4_K_M | ~205 MB | Smallest here, most aggressive quantization, best for constrained devices |

(Sizes are approximate — check actual file sizes in the repo. 350M params ≈ 1.5x the size of the 230M variants.)

Usage

llama.cpp

./llama-cli -m lfm2.5-350m-python-math-Q5_K_S.gguf -t 8 --temperature 0.5 --top-p 0.9 --top-k 50 --min-p 0.05 --repeat-penalty 1.1

Ollama

ollama run hf.co/hauser458original/lfm2.5-350m-python-math-GGUF:Q5_K_S

LM Studio

Search for hauser458original/lfm2.5-350m-python-math-GGUF in the LM Studio model browser, or download a .gguf file directly and load it manually.

Which quant should I use?

• Q4_K_M: smallest footprint, best for very constrained devices. Some quality loss vs. higher quants.
• Q5_K_S / Q5_K_M: recommended default for most laptop/desktop CPU inference. Best speed/quality tradeoff.
• Q8_0: near-lossless, use if you have the RAM/storage headroom.
• F16: full precision GGUF, only needed if you plan to re-quantize yourself.

License

Inherits the LFM Open License v1.0 from the base model.