GGUF
conversational
How to use from
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 Odontome/LFM2.5-1.2B-Elm:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf Odontome/LFM2.5-1.2B-Elm:Q4_K_M
Install from WinGet (Windows)
winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf Odontome/LFM2.5-1.2B-Elm:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf Odontome/LFM2.5-1.2B-Elm: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 Odontome/LFM2.5-1.2B-Elm:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama cli -hf Odontome/LFM2.5-1.2B-Elm:Q4_K_M
Use Docker
docker model run hf.co/Odontome/LFM2.5-1.2B-Elm:Q4_K_M
Quick Links

banner

LFM2.5-1.2B-Elm

Model Description

LFM2.5-1.2B-Elm is a fine-tuned version of LiquidAI/LFM2.5-1.2B-Thinking.

The agentic capabilities have been specifically enhanced for the Hermes-Agent harness. This model was trained using the following datasets:

  • kai-os/carnice-glm5-hermes-traces
  • lambda/hermes-agent-reasoning-traces

Performance & Benchmarks

This model is optimized for speed and efficiency. Inference speeds were measured using llama.cpp (CUDA v2.11.0) within LM Studio:

Hardware Tokens/s Notes
RTX 3090 ~506 Tok/s Blazing fast inference
RTX 4060 Laptop ~250 Tok/s Excellent portability performance
Downloads last month
27
GGUF
Model size
1B params
Architecture
lfm2
Hardware compatibility
Log In to add your hardware

4-bit

Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for Odontome/LFM2.5-1.2B-Elm

Quantized
(58)
this model

Datasets used to train Odontome/LFM2.5-1.2B-Elm