[GH-ISSUE #15983] Model Proposal: CAJAL-9B-P2PCLAW — Scientific Paper Generation Modelfile #87865

Closed
opened 2026-05-10 06:28:07 -05:00 by GiteaMirror · 0 comments
Owner

Originally created by @Agnuxo1 on GitHub (May 5, 2026).
Original GitHub issue: https://github.com/ollama/ollama/issues/15983

CAJAL-9B-P2PCLAW Model Proposal

Model Overview

Name: CAJAL-9B-P2PCLAW
Base Model: Qwen3.6-9B-Instruct
Purpose: Local scientific paper generation (trained on 50+ peer-reviewed papers)
Context Window: 32K tokens
Special Tokens: [ABSTRACT], [INTRO], [METHODS], [RESULTS], [CONCLUSIONS]
GGUF Sizes: Q4_K_M (~4.5GB), Q5_K_M (~5.5GB), Q8_0 (~9.1GB)

Why add CAJAL to the Ollama library?

  1. Unique niche — First Ollama model specialized for scientific/academic writing
  2. Local-first — No API keys, runs entirely offline
  3. Lightweight — 4.5GB Q4_K_M fits consumer GPUs
  4. Paper-aware — Understands IMRaD structure, LaTeX, citations
  5. Open source — Apache 2.0, full weights on HuggingFace

Usage Example

ollama pull Agnuxo/CAJAL-9B-P2PCLAW
ollama run CAJAL-9B-P2PCLAW

# Generate paper abstract
>>> Write a scientific abstract about neural networks in climate modeling
[ABSTRACT] This study investigates the application of deep neural networks...

Modelfile

FROM ./cajal-9b-q4_k_m.gguf
PARAMETER temperature 0.7
PARAMETER num_ctx 32768
SYSTEM You are CAJAL, a specialized scientific paper generation assistant. Use [ABSTRACT], [INTRO], [METHODS], [RESULTS], and [CONCLUSIONS] tokens for structured output.

Would love to collaborate on getting CAJAL into the official Ollama library! 🚀

Originally created by @Agnuxo1 on GitHub (May 5, 2026). Original GitHub issue: https://github.com/ollama/ollama/issues/15983 # CAJAL-9B-P2PCLAW Model Proposal ## Model Overview **Name:** CAJAL-9B-P2PCLAW **Base Model:** Qwen3.6-9B-Instruct **Purpose:** Local scientific paper generation (trained on 50+ peer-reviewed papers) **Context Window:** 32K tokens **Special Tokens:** [ABSTRACT], [INTRO], [METHODS], [RESULTS], [CONCLUSIONS] **GGUF Sizes:** Q4_K_M (~4.5GB), Q5_K_M (~5.5GB), Q8_0 (~9.1GB) ## Why add CAJAL to the Ollama library? 1. **Unique niche** — First Ollama model specialized for scientific/academic writing 2. **Local-first** — No API keys, runs entirely offline 3. **Lightweight** — 4.5GB Q4_K_M fits consumer GPUs 4. **Paper-aware** — Understands IMRaD structure, LaTeX, citations 5. **Open source** — Apache 2.0, full weights on HuggingFace ## Usage Example ```bash ollama pull Agnuxo/CAJAL-9B-P2PCLAW ollama run CAJAL-9B-P2PCLAW # Generate paper abstract >>> Write a scientific abstract about neural networks in climate modeling [ABSTRACT] This study investigates the application of deep neural networks... ``` ## Links - 🤗 HuggingFace: https://huggingface.co/Agnuxo/CAJAL-9B-P2PCLAW - 📁 GitHub: https://github.com/Agnuxo1/CAJAL - 🌐 Project: https://www.p2pclaw.com/silicon ## Modelfile ```dockerfile FROM ./cajal-9b-q4_k_m.gguf PARAMETER temperature 0.7 PARAMETER num_ctx 32768 SYSTEM You are CAJAL, a specialized scientific paper generation assistant. Use [ABSTRACT], [INTRO], [METHODS], [RESULTS], and [CONCLUSIONS] tokens for structured output. ``` Would love to collaborate on getting CAJAL into the official Ollama library! 🚀
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#87865