LFM3-VL-2B-GGUF
GGUF version of LFM3-VL-2B for local inference on CPUs and GPUs.
Find architecture and usage details in the original model card.
How to run LFM3-VL-2B
Recommended sampling: temperature 0.1, top_k 50, top_p 1.0, min_p 0.0, repeat penalty 1.1.
- Llama.app (Use this model > Llama in the menu), for macOS and Windows: image and video input. Requires llama.cpp version v0.8.0 or newer.
- LM Studio (Use this model > LM Studio in the menu), for macOS, Windows and Linux: image input. Requires LM Studio with llama.cpp engine 2.65.1 or newer.
- Ollama
ollama run lfm3-vl, for macOS: image input. Requires Ollama version v0.40.12 or newer.
NVFP4 GGUF
Unlike the other files in this repository, which are converted from LFM3-VL-2B,
LFM3-VL-2B-NVFP4.gguf is converted from a separate repository,
LiquidAI/LFM3-VL-2B-NVFP4, a checkpoint quantized to NVFP4 with NVIDIA Model Optimizer.
The language model's feed-forward weights keep their NVFP4 quantization from that checkpoint, all other weights are Q8_0.
The vision encoder is not quantized in that checkpoint, so the mmproj files in this repository are used unchanged.
On NVIDIA Blackwell GPUs (RTX 50 series, DGX Spark) the NVFP4 layers run on FP4 tensor cores.
[!NOTE] Supported in llama.cpp b12088 and newer (#31608). Older builds load the file but produce incorrect output.
Example usage with llama.cpp:
llama-mtmd-cli -hf LiquidAI/LFM3-VL-2B-GGUF:NVFP4 \
--image image.jpg -p "Describe this image." \
--temp 0.1 --top-k 50 --top-p 1.0 --min-p 0.0 --repeat-penalty 1.1
QAD Q4_0 GGUF
The Q4_0 build in this repository is the Quantization-Aware Distillation (QAD) checkpoint,
LFM3-VL-2B-QAD-Q4_0.gguf.
It was trained to recover the accuracy lost to Q4_0 quantization.