MiniCPM5-1B-sft — computational-chemistry assistant (merged, Ollama-ready)
Merged FP16 weights = base openbmb/MiniCPM5-1B-sft + the completion-only SFT adapter, from https://github.com/anisiraj/llm_kickstart_repo (handbook.html → 🧪 Case Study).
Llama-architecture → Ollama imports it directly. Measured: 0.689 GB at Q4_K_M; completion ppl →3.64, recall 46%
# quantize to Q4_K_M (~0.7 GB) and serve locally
ollama create chem-minicpm5 -q q4_K_M -f Modelfile
ollama run chem-minicpm5 "What is the Hartree-Fock method?"
- Downloads last month
- 13
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support