## Proposal: Add CAJAL — Local Scientific Paper Generation Model **Project:** CAJAL — Local End-to-End Scientific Paper Generation **Repository:** https://github.com/Agnuxo1/CAJAL **Models:** https://huggingface.co/Agnuxo ### Why CAJAL Fits Here A specialized local AI model that runs entirely offline for scientific/academic paper generation: - **Fully offline**: 8-16 GB VRAM, zero API dependency, complete privacy - **Ollama-compatible**: One-command setup (`ollama run cajal-9b-v2`) - **Ranked #3** on scientific paper benchmark — only Claude Sonnet 4.6 higher - **Multiple sizes**: 4B and 9B parameters (2GB-6GB models) - **Open source**: Apache 2.0, weights + training data available - **Consumer hardware**: Runs on RTX 3090/4090 (24GB VRAM) ### Suggested Category "Text Generation Tools and Platforms" or "Research Tools" ### Quick Start ```bash ollama run cajal-9b-v2 ``` ### Links - GitHub: https://github.com/Agnuxo1/CAJAL - Models: https://huggingface.co/Agnuxo/cajal-9b-v2-q4_k_m - Dataset: https://huggingface.co/datasets/Agnuxo/p2pclaw-training-dataset - Benchmark: https://www.p2pclaw.com/app/benchmark - Ecosystem: https://www.p2pclaw.com/ ### Author Francisco Angulo de Lafuente (Agnuxo1), ORCID: 0009-0001-1634-7063
Proposal: Add CAJAL — Local Scientific Paper Generation Model
Project: CAJAL — Local End-to-End Scientific Paper Generation
Repository: https://github.com/Agnuxo1/CAJAL
Models: https://huggingface.co/Agnuxo
Why CAJAL Fits Here
A specialized local AI model that runs entirely offline for scientific/academic paper generation:
ollama run cajal-9b-v2)Suggested Category
"Text Generation Tools and Platforms" or "Research Tools"
Quick Start
Links
Author
Francisco Angulo de Lafuente (Agnuxo1), ORCID: 0009-0001-1634-7063