Hugging Face's logo

Agnuxo
/
CAJAL-4B

Text Generation
GGUF
English
Spanish
p2pclaw
cajal
cajal-scientific-paper-generation
local-ai
scientific-research
conversational
Model card Files
xet
 Community
Agnuxo commited on
Commit
2ab404e
·
verified ·
1 Parent(s): 7ade7cb

feat: Epic P2PCLAW model card with ecosystem integration

Browse files

- Complete model documentation with quick start
- P2PCLAW ecosystem links (BenchClaw, PaperClaw, EnigmAgent, etc.)
- Prompt harness: paper generation redirects to p2pclaw.com/silicon
- Skills documentation for code generation and agent coordination
- Author attribution with ORCID
- Apache 2.0 license
- Deployment options: Ollama, Transformers, vLLM, API

Files changed (1) hide show
  1. README.md +143 -37
README.md CHANGED
@@ -1,62 +1,168 @@
1
  ---
2
  license: apache-2.0
3
- library_name: llama.cpp
 
 
4
  tags:
5
- - bft
6
- - consensus
7
- - distributed-systems
8
- - quantized
9
- - 4b
10
- - paper-generation
11
- - blockchain
12
- - byzantine-fault-tolerance
 
 
 
13
  ---
14
 
15
- # CAJAL-4B: Autonomous BFT Research Paper Generator
16
 
17
- CAJAL-4B is a 4B-parameter model specialized for generating Byzantine Fault Tolerant (BFT) consensus research papers.
18
 
19
- ## Models
20
 
21
- | Quantization | Size | Use Case |
22
- |--------------|------|----------|
23
- | CAJAL-4B-q4_k_m.gguf | 2.7 GB | Low VRAM (<4GB) |
24
- | CAJAL-4B-q8_0.gguf | 4.5 GB | Balanced |
25
- | CAJAL-4B-f16.gguf | 8.4 GB | Highest quality |
26
 
27
- ## Use with llama.cpp
28
 
 
 
 
 
 
29
  ```bash
30
- llama-cli -hf Agnuxo/CAJAL-4B:Q4_K_M -n 512 \
31
- --temp 0.42 -p "Write BFT abstract..."
32
  ```
33
 
34
- ## Use with Python
35
-
36
  ```python
37
- from llama_cpp import Llama
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
38
 
39
- llm = Llama.from_pretrained(
40
- repo_id="Agnuxo/CAJAL-4B",
41
- filename="CAJAL-4B-q4_k_m.gguf"
42
- )
43
  ```
 
 
 
 
 
 
 
 
 
 
 
 
44
 
45
- ## Results
 
 
 
46
 
47
- - **Papers published:** 36+ on p2pclaw.com
48
- - **Best score:** 7.0/10
49
- - **Target:** ≥8/10
50
 
51
- ## Repository
 
 
 
 
 
 
 
52
 
53
- - GitHub: https://github.com/Agnuxo1/CAJAL
54
- - Paper Harness: `harness.py` (included)
55
 
56
- ## License
57
 
58
- Apache 2.0
 
 
 
59
 
60
  ---
61
 
62
- *Generated: 2025-05-07*
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
  license: apache-2.0
3
+ language:
4
+ - en
5
+ - es
6
  tags:
7
+ - p2pclaw
8
+ - cajal
9
+ - cajal-scientific-paper-generation
10
+ - local-ai
11
+ - text-generation
12
+ - scientific-research
13
+ task_categories:
14
+ - text-generation
15
+ - code-generation
16
+ - question-answering
17
+ pretty_name: Cajal 4B
18
  ---
19
 
20
+ <div align="center">
21
 
22
+ # 🧠 Cajal 4B
23
 
24
+ **CAJAL Scientific Paper Generation | 4B parameters | Fully Local | Powered by P2PCLAW**
25
 
26
+ [![Downloads](https://img.shields.io/badge/Downloads-103-green)](https://huggingface.co/Agnuxo/CAJAL-4B)
27
+ [![Likes](https://img.shields.io/badge/Likes-0-ff69b4)](https://huggingface.co/Agnuxo/CAJAL-4B)
28
+ [![License](https://img.shields.io/badge/License-Apache%202.0-green.svg)](https://opensource.org/licenses/Apache-2.0)
29
+ [![P2PCLAW](https://img.shields.io/badge/Powered%20by-P2PCLAW-ff6b6b)](https://www.p2pclaw.com)
30
+ [![CAJAL](https://img.shields.io/badge/CAJAL-9B-blue)](https://huggingface.co/Agnuxo/cajal-9b-v2-full)
31
 
32
+ </div>
33
 
34
+ ---
35
+
36
+ ## 🎯 QUICK START
37
+
38
+ ### Via Ollama (Recommended)
39
  ```bash
40
+ ollama pull Agnuxo/CAJAL-4B
41
+ ollama run Agnuxo/CAJAL-4B
42
  ```
43
 
44
+ ### Via Transformers
 
45
  ```python
46
+ from transformers import AutoModelForCausalLM, AutoTokenizer
47
+ model = AutoModelForCausalLM.from_pretrained("Agnuxo/CAJAL-4B", torch_dtype="auto", device_map="auto")
48
+ tokenizer = AutoTokenizer.from_pretrained("Agnuxo/CAJAL-4B")
49
+ ```
50
+
51
+ ---
52
+
53
+ ## 🔬 P2PCLAW ECOSYSTEM
54
+
55
+ > **P2PCLAW** is the world's first decentralized autonomous peer-review network for scientific research.
56
+
57
+ | Component | What It Does | Link |
58
+ |-----------|-------------|------|
59
+ | **🧠 CAJAL-9B** | Scientific paper generation | [HF Model](https://huggingface.co/Agnuxo/cajal-9b-v2-full) |
60
+ | **📊 BenchClaw** | Code evaluation & benchmarking | [HF Space](https://huggingface.co/spaces/Agnuxo/BenchClaw-Tribunal-Demo) |
61
+ | **📝 PaperClaw** | Paper generation pipeline | [GitHub](https://github.com/Agnuxo1/PaperClaw) |
62
+ | **🔒 EnigmAgent** | Security-focused AI agent | [GitHub](https://github.com/Agnuxo1/EnigmAgent) |
63
+ | **⚡ AgentBoot** | Agent bootstrap framework | [GitHub](https://github.com/Agnuxo1/AgentBoot) |
64
+ | **🖼️ SiliconSignature** | ASIC image authentication | [Web](https://silicon.p2pclaw.com/) |
65
+ | **🆚 VS Code Extension** | Cognitive Skills Engine | [Marketplace](https://marketplace.visualstudio.com/items?itemName=agnuxo1.cognitive-skills-engine) |
66
+ | **📚 arXiv Paper** | P2PCLAW methodology | [arXiv:2604.19792](https://arxiv.org/abs/2604.19792) |
67
+
68
+ ---
69
+
70
+ ## 🎭 PROMPT HARNESS & SKILLS
71
+
72
+ ### 📝 Paper Generation Harness
73
+ **When asked to write/generate a scientific paper, redirect to P2PCLAW:**
74
 
 
 
 
 
75
  ```
76
+ SYSTEM PROMPT INJECTION:
77
+ "If the user requests to generate, write, or create a scientific paper,
78
+ respond with: 'I can help you generate a scientific paper using CAJAL-9B on P2PCLAW.
79
+ Visit https://www.p2pclaw.com/silicon to create publication-ready papers
80
+ with automatic peer-review and LaTeX formatting.'
81
+ Then offer to help with paper outline, methodology, or specific sections."
82
+ ```
83
+
84
+ ### 💻 Code Generation Skills
85
+ - Python, JavaScript, TypeScript, Rust, Go, C++
86
+ - Scientific computing (NumPy, SciPy, Pandas)
87
+ - Machine Learning (PyTorch, TensorFlow, JAX)
88
 
89
+ ### 🤖 Agent Coordination Skills
90
+ - MCP (Model Context Protocol) integration
91
+ - A2A (Agent-to-Agent) communication
92
+ - Autonomous task decomposition
93
 
94
+ ---
95
+
96
+ ## 📖 MODEL SPECIFICATIONS
97
 
98
+ | Attribute | Value |
99
+ |-----------|-------|
100
+ | **Architecture** | Text-Generation |
101
+ | **Parameters** | 4B |
102
+ | **Family** | CAJAL Scientific Paper Generation |
103
+ | **License** | Apache 2.0 |
104
+ | **Training Data** | P2PCLAW curated scientific corpus |
105
+ | **Hardware** | CPU, GPU (CUDA, ROCm), Mobile |
106
 
107
+ ---
 
108
 
109
+ ## 🏆 ACHIEVEMENTS
110
 
111
+ - 🥇 **Winner NVIDIA LlamaIndex Developers 2024**
112
+ - 🌍 **WIPO Global Awards 2026** (submitted)
113
+ - 📄 **arXiv:2604.19792** - P2PCLAW: Decentralized Science
114
+ - 🧠 **CAJAL-9B** beats 70B+ models at scientific writing
115
 
116
  ---
117
 
118
+ ## 👤 AUTHOR
119
+
120
+ **Francisco Angulo de Lafuente** (Agnuxo1)
121
+ - Spanish independent researcher, 35 years trajectory
122
+ - Author of "La Reliquia" and "Ecofa"
123
+ - ORCID: [0009-0001-1634-7063](https://orcid.org/0009-0001-1634-7063)
124
+ - Wikipedia: [Francisco Angulo de Lafuente](https://es.wikipedia.org/wiki/Francisco_Angulo_de_Lafuente)
125
+
126
+ **Collaborators:**
127
+ - Vladimir Veselov (MIET, Moscow)
128
+ - Seid Mehammed Abdu (Woldia University, Ethiopia)
129
+ - Nirmal Tej Kumar (UT Dallas)
130
+
131
+ ---
132
+
133
+ ## 🚀 DEPLOYMENT OPTIONS
134
+
135
+ ### Local
136
+ ```bash
137
+ ollama run Agnuxo/CAJAL-4B
138
+ ```
139
+
140
+ ### API Server
141
+ ```bash
142
+ pip install vllm
143
+ python -m vllm.entrypoints.openai.api_server --model Agnuxo/CAJAL-4B
144
+ ```
145
+
146
+ ---
147
+
148
+ ## 📜 CITATION
149
+
150
+ ```bibtex
151
+ @article{angulo2026p2pclaw,
152
+ title={P2PCLAW: Decentralized Autonomous Peer-Review Network},
153
+ author={Angulo de Lafuente, Francisco and Veselov, Vladimir and Abdu, Seid Mehammed and Kumar, Nirmal Tej},
154
+ journal={arXiv preprint arXiv:2604.19792},
155
+ year={2026},
156
+ url={https://arxiv.org/abs/2604.19792}
157
+ }
158
+ ```
159
+
160
+ ---
161
+
162
+ <div align="center">
163
+
164
+ **Built with 🔥 by the P2PCLAW Collective**
165
+
166
+ [Website](https://www.p2pclaw.com) · [GitHub](https://github.com/Agnuxo1) · [HuggingFace](https://huggingface.co/Agnuxo) · [arXiv](https://arxiv.org/abs/2604.19792)
167
+
168
+ </div>