LiquidAI/LFM2.5-1.2B-Instruct
RepoLFM-2.5-1.2B-Instruct is a 1.2-billion-parameter, instruction-tuned language model released by the Liquid AI research group on Hugging Face. It provides an open-source, lightweight alternative for developers who need an LLM that can run on modest hardware while still following user prompts accurately.
article
3 storys
calendar_today
First: 2026-02-20
update
Last: 2026-08-21
Stories
Completed digest stories linked to this service.
-
DSpark lands in llama.cpp: 2–3x faster on‑device decode for LFM2.52026-08-21LiquidAI’s LFM2.5 draft models add DSpark speculative decoding with upstream llama.cpp support, delivering big...
-
Post-training is doing the heavy lifting: GLM-5.3 jumps without new pretraining2026-08-20Z.ai's GLM-5.3 delivered big coding and security gains from post-training alone, signaling a shift from retrai...
-
Practical LLM efficiency: Magma optimizer, Unsloth on HF Jobs, and NVLink realit...2026-02-20A new wave of efficiency wins—masked optimizers, free small‑model fine‑tuning, and faster GPU interconnects—ca...