README.md

metadata

title: sci_paper_llm
emoji: 🔬
colorFrom: blue
colorTo: indigo
sdk: docker
app_port: 7860
pinned: false
license: mit

sci_paper_llm — live demo

Build open-access scientific-paper datasets from OpenAlex and query them with a local LLM, running entirely inside this Space.

Source code: https://github.com/RebornX10/rebornx10.github.io

Free CPU Space caveats: the LLM runs on CPU with a small model (qwen2.5:0.5b by default), so answers are slow and the model is pulled on cold start. For full speed, use a GPU Space or run locally with Docker Compose. Change the model with the OLLAMA_MODEL Space variable.