costinstroie/lfm-medical-1.2b-GGUF

lfm-medical-1.2b-GGUF

GGUF conversion of burtenshaw/unsloth-lfm-medical-lr-5e-5-20260211b, a LoRA (r=16, alpha=16, SFT with TRL/Unsloth) fine-tune of unsloth/LFM2.5-1.2B-Instruct for medical Q&A.

The adapter was merged into the base weights (fp32, stored as bf16) and converted with llama.cpp (b8733). All credit for the fine-tune goes to the original author; this repo only repackages it.

Usage

The embedded chat template is the stock LFM2.5 template.

llama-cli -m lfm-medical-1.2b-Q8_0.gguf

The model typically begins its reply with a <reasoning>…</reasoning> block before the answer; strip it if you only want the final text.

Limitations

The source repository has no model card, so the training data is undocumented. Not for clinical use — outputs may be wrong and must be reviewed by a qualified professional.

License

Derived from LFM2.5-1.2B-Instruct and subject to the LFM Open License v1.0 (see LICENSE), including its annual-revenue threshold of US$10M for commercial use. Read the license before using this model commercially.