OpenCLAW-P2P v6.0: Resilient Multi-Layer Persistence, Live Reference Verification, and Production-Scale Evaluation of Decentralized AI Peer Review
Paper β’ 2604.19792 β’ Published β’ 1
id stringlengths 13 117 | repo stringclasses 8
values | project stringclasses 1
value | author stringclasses 1
value | author_orcid stringclasses 1
value | filename stringlengths 5 44 | path stringlengths 5 106 | extension stringclasses 36
values | content stringlengths 61 68.4M | size int64 61 68.4M | lines int64 2 13.6k | type stringclasses 10
values | source_url stringclasses 7
values | license stringclasses 1
value | timestamp stringlengths 19 26 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
CAJAL/test_import.py | CAJAL | P2PCLAW | Francisco Angulo de Lafuente | 0009-0001-1634-7063 | test_import.py | test_import.py | .py | import os
os.environ["UNSLOTH_COMPILE_DISABLE"] = "1"
import sys
# Try importing unsloth piece by piece to find the crash
modules = [
"unsloth._utils",
"unsloth.models",
"unsloth.save",
"unsloth.chat_templates",
]
for mod in modules:
try:
__import__(mod)
print(f" {mod}: OK", flus... | 772 | 30 | python | https://github.com/Agnuxo1/CAJAL | Apache-2.0 | 2026-05-10T21:03:37.903400 |
CAJAL/README.ru.md | CAJAL | P2PCLAW | Francisco Angulo de Lafuente | 0009-0001-1634-7063 | README.ru.md | README.ru.md | .md | # π§ CAJAL
> **ΠΠΎΠ³Π½ΠΈΡΠΈΠ²Π½ΡΠΉ ΡΠ»ΠΎΠΉ Π΄Π»Ρ Π½Π°ΠΏΠΈΡΠ°Π½ΠΈΡ Π°ΠΊΠ°Π΄Π΅ΠΌΠΈΡΠ΅ΡΠΊΠΈΡ
ΠΆΡΡΠ½Π°Π»ΠΎΠ²** β ΠΠ΅Π½Π΅ΡΠΈΡΡΠΉΡΠ΅ Π½Π°ΡΡΠ½ΡΠ΅ ΡΡΠ°ΡΡΠΈ, Π³ΠΎΡΠΎΠ²ΡΠ΅ ΠΊ ΠΏΡΠ±Π»ΠΈΠΊΠ°ΡΠΈΠΈ, Π»ΠΎΠΊΠ°Π»ΡΠ½ΠΎ, Π±Π΅ΡΠΏΠ»Π°ΡΠ½ΠΎ ΠΈ Π±Π΅Π· Π·Π°Π²ΠΈΡΠΈΠΌΠΎΡΡΠΈ ΠΎΡ ΠΎΠ±Π»Π°ΠΊΠ°.
[](https://pypi.org/project/cajal-p2pclaw/)
[, founder of the P2PCLAW Research Network. We build open-source tools for decentralized scientific research.
**What I'm offering (completely free, no strings... | 1,633 | 36 | documentation | https://github.com/Agnuxo1/CAJAL | Apache-2.0 | 2026-05-10T21:03:37.903557 |
CAJAL/Modelfile | CAJAL | P2PCLAW | Francisco Angulo de Lafuente | 0009-0001-1634-7063 | Modelfile | Modelfile | # CAJAL-9B Modelfile for Ollama
# Local Scientific Paper Generation Agent
# Part of P2PCLAW Ecosystem
FROM ./cajal-9b-q4_k_m.gguf
# Model parameters optimized for scientific paper generation
PARAMETER temperature 0.3
PARAMETER top_p 0.8
PARAMETER top_k 40
PARAMETER repeat_penalty 1.1
PARAMETER num_ctx 32768
PARAMETER... | 1,768 | 63 | text | https://github.com/Agnuxo1/CAJAL | Apache-2.0 | 2026-05-10T21:03:37.903618 | |
CAJAL/check_model.py | CAJAL | P2PCLAW | Francisco Angulo de Lafuente | 0009-0001-1634-7063 | check_model.py | check_model.py | .py | from transformers import AutoConfig
c = AutoConfig.from_pretrained(r'D:\PROJECTS\CAJAL\Modelos_originales\Qwen3.5-4B', trust_remote_code=True)
print(f'model_type: {c.model_type}')
print(f'architectures: {c.architectures}')
print(f'num_hidden_layers: {getattr(c, "num_hidden_layers", "N/A")}')
print(f'hidden_size: {getat... | 406 | 7 | python | https://github.com/Agnuxo1/CAJAL | Apache-2.0 | 2026-05-10T21:03:37.903691 |
CAJAL/MODEL_CARD.md | CAJAL | P2PCLAW | Francisco Angulo de Lafuente | 0009-0001-1634-7063 | MODEL_CARD.md | MODEL_CARD.md | .md | ---
language:
- en
- es
- zh
- de
- fr
license: apache-2.0
library_name: transformers
tags:
- ollama
- gguf
- transformers
- safetensors
- qwen3.5
- causal-lm
- lora
- qlora
- text-generation
- conversational
- agent
- scientific-research
- peer-to-peer
- crypto-law
- p2pclaw
- fine-tuned
base_model: Qwen/Qwen3.5-4B
pi... | 9,338 | 263 | documentation | https://github.com/Agnuxo1/CAJAL | Apache-2.0 | 2026-05-10T21:03:37.904004 |
CAJAL/INTEGRATIONS.md | CAJAL | P2PCLAW | Francisco Angulo de Lafuente | 0009-0001-1634-7063 | INTEGRATIONS.md | INTEGRATIONS.md | .md | # CAJAL-4B Integration Ecosystem
> **Universal integration layer for CAJAL-4B** - Deploy the world's first scientific intelligence model for P2P systems across any platform.
[](https://pypi.org/project/cajal-cli/)
[![HuggingFace](https://img.shields.io/badge/HF-Agnu... | 6,257 | 244 | documentation | https://github.com/Agnuxo1/CAJAL | Apache-2.0 | 2026-05-10T21:03:37.904120 |
CAJAL/test_9b_load.py | CAJAL | P2PCLAW | Francisco Angulo de Lafuente | 0009-0001-1634-7063 | test_9b_load.py | test_9b_load.py | .py | import sys
import io
if sys.platform == "win32":
sys.stdout = io.TextIOWrapper(sys.stdout.buffer, encoding='utf-8')
sys.stderr = io.TextIOWrapper(sys.stderr.buffer, encoding='utf-8')
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer
import gc
model_path = r"D:\PROJECTS\CAJAL\Modelos_or... | 1,488 | 48 | python | https://github.com/Agnuxo1/CAJAL | Apache-2.0 | 2026-05-10T21:03:37.904180 |
CAJAL/install_wsl2.bat | CAJAL | P2PCLAW | Francisco Angulo de Lafuente | 0009-0001-1634-7063 | install_wsl2.bat | install_wsl2.bat | .bat | @echo off
echo ========================================
echo Instalando WSL2 para CAJAL Training
echo ========================================
echo Habilitando features de Windows...
dism.exe /online /enable-feature /featurename:Microsoft-Windows-Subsystem-Linux /all /norestart
dism.exe /online /enable-feature /feature... | 661 | 16 | text | https://github.com/Agnuxo1/CAJAL | Apache-2.0 | 2026-05-10T21:03:37.904230 |
CAJAL/plan_cajal.md | CAJAL | P2PCLAW | Francisco Angulo de Lafuente | 0009-0001-1634-7063 | plan_cajal.md | plan_cajal.md | .md | # Plan Ampliado: CAJAL-4B β Dataset Ecosystem Completo
## Contexto Actualizado
- **Nombre del modelo**: CAJAL + parΓ‘metros (CAJAL-4B, CAJAL-8B, etc.)
- **Dataset base**: ~670 papers de P2PCLAW
- **Dataset ampliado**: Papers + Repositorios + Skills + Archivos locales + Recursos externos
## Fuentes de Conocimiento a In... | 3,253 | 70 | documentation | https://github.com/Agnuxo1/CAJAL | Apache-2.0 | 2026-05-10T21:03:37.904354 |
CAJAL/start_training_hidden.vbs | CAJAL | P2PCLAW | Francisco Angulo de Lafuente | 0009-0001-1634-7063 | start_training_hidden.vbs | start_training_hidden.vbs | .vbs | Set WshShell = CreateObject("WScript.Shell")
WshShell.Run "cmd /c cd /d D:\PROJECTS\CAJAL && python scripts\train_cajal_4b.py --model-path ""D:\PROJECTS\CAJAL\Modelos_originales\Qwen3.5-4B"" --dataset ""D:\PROJECTS\CAJAL\cajal_dataset.jsonl"" --output-dir ""D:\PROJECTS\CAJAL\outputs\CAJAL-4B"" --output-name CAJAL-4B --... | 573 | 2 | text | https://github.com/Agnuxo1/CAJAL | Apache-2.0 | 2026-05-10T21:03:37.904428 |
CAJAL/ollama-modelfile | CAJAL | P2PCLAW | Francisco Angulo de Lafuente | 0009-0001-1634-7063 | ollama-modelfile | ollama-modelfile | # CAJAL Modelfile for Ollama
FROM llama3.1
SYSTEM """You are CAJAL, an autonomous scientific paper generator with built-in peer review tribunal.
## Paper Generation Protocol
When the user requests a paper:
1. **Draft Generation**
- Generate 7 sections: Abstract, Introduction, Related Work, Methodology, Results,... | 1,336 | 51 | text | https://github.com/Agnuxo1/CAJAL | Apache-2.0 | 2026-05-10T21:03:37.904479 |
End of preview. Expand in Data Studio
YAML Metadata Warning:The task_categories "code-generation" is not in the official list: text-classification, token-classification, table-question-answering, question-answering, zero-shot-classification, translation, summarization, feature-extraction, text-generation, fill-mask, sentence-similarity, text-to-speech, text-to-audio, automatic-speech-recognition, audio-to-audio, audio-classification, audio-text-to-text, voice-activity-detection, depth-estimation, image-classification, object-detection, image-segmentation, text-to-image, image-to-text, image-to-image, image-to-video, unconditional-image-generation, video-classification, reinforcement-learning, robotics, tabular-classification, tabular-regression, tabular-to-text, table-to-text, multiple-choice, text-ranking, text-retrieval, time-series-forecasting, text-to-video, image-text-to-text, image-text-to-image, image-text-to-video, visual-question-answering, document-question-answering, zero-shot-image-classification, graph-ml, mask-generation, zero-shot-object-detection, text-to-3d, image-to-3d, image-feature-extraction, video-text-to-text, keypoint-detection, visual-document-retrieval, any-to-any, video-to-video, other
𧬠P2PCLAW Ecosystem β Complete Training Dataset
638 files. 161 MB. The entire knowledge base of Francisco Angulo de Lafuente (Agnuxo1) and the P2PCLAW decentralized research network.
π What's Inside
This dataset contains the complete intellectual output of Francisco Angulo de Lafuente's 35-year research trajectory, packaged for training the next generation of scientific AI models.
| Category | Files | Description |
|---|---|---|
| Documentation | 148 | READMEs, technical docs, specifications |
| Python Code | 132 | Models, agents, benchmarks, infrastructure |
| JavaScript | 86 | Web UIs, demos, browser extensions |
| Configuration | 99 | JSON configs, YAML pipelines, setup files |
| Web Assets | 32 | HTML, CSS for web applications |
| Shell Scripts | 19 | Deployment, automation, devops |
| Academic/Math | 22 | Papers, equations, formal proofs (Lean 4) |
| Biography | 1 | Complete author profile and history |
π§ Projects Included
| Project | Files | Description | Links |
|---|---|---|---|
| CAJAL-9B | 302 | Scientific paper generation model | Model Β· Paper |
| P2PCLAW | 2 | Decentralized research network | Website |
| EnigmAgent | 307 | Security-focused autonomous agent | Demo |
| SiliconSignature | 14 | ASIC-based image authentication | Web |
| Winner NVIDIA 2024 | 10 | LlamaIndex award-winning project | GitHub |
| Francisco Angulo | 2 | Personal website and biography | Wiki |
π€ About the Author
Francisco Angulo de Lafuente (Agnuxo1) is a Spanish independent researcher, author, and visionary who has spent 35 years bridging literature and technology.
Literary Works
- "La Reliquia" β Novel that introduced decentralized knowledge concepts 20+ years ago
- "Ecofa" β Novel exploring biofuels and sustainability
Research Impact
- P2PCLAW: First decentralized autonomous peer-review network
- CAJAL-9B: 9B model that beats 70B+ models at scientific writing
- SiliconSignature: Patent-ready ASIC-based image authentication
- 14 autonomous agents coordinating scientific research
Academic Network
- Vladimir Veselov β MIET, Moscow
- Seid Mehammed Abdu β Woldia University, Ethiopia
- Nirmal Tej Kumar β UT Dallas
Recognition
- Winner NVIDIA LlamaIndex Developers 2024
- WIPO Global Awards 2026 (submitted)
- ORCID: 0009-0001-1634-7063
π Usage
For LLM Training
from datasets import load_dataset
ds = load_dataset("Agnuxo/p2pclaw-ecosystem-dataset", split="train")
# Train on scientific code generation
code_samples = ds.filter(lambda x: x["type"] == "python")
# Train on technical documentation
docs = ds.filter(lambda x: x["type"] == "documentation")
# Train on the author's complete knowledge base
all_content = ds.filter(lambda x: x["size"] > 500)
For RAG (Retrieval-Augmented Generation)
# Index all P2PCLAW knowledge for retrieval
from sentence_transformers import SentenceTransformer
model = SentenceTransformer('all-MiniLM-L6-v2')
embeddings = model.encode(ds["content"])
# Use with vector DB for question answering about P2PCLAW
π¦ File Structure
p2pclaw-ecosystem-dataset/
βββ CAJAL.jsonl # 302 records β Paper generator codebase
βββ CAJAL-chunk1.jsonl # Small records (< 10MB each)
βββ CAJAL-chunk2.jsonl.gz # Large records (13MB compressed)
βββ CAJAL-chunk3.jsonl.gz # Medium records (7MB compressed)
βββ CAJAL-chunk4.jsonl.gz # Medium records (8MB compressed)
βββ CAJAL-chunk5.jsonl.gz # Medium records (10MB compressed)
βββ CAJAL-chunk6.jsonl # Small records (< 10MB)
βββ EnigmAgent.jsonl # 307 records β Security agent
βββ P2PCLAW.jsonl # 2 records β Network infrastructure
βββ SiliconSignature.jsonl # 14 records β Image authentication
βββ Winner-Nvidia-2024.jsonl # 10 records β Award project
βββ Francisco-Angulo.jsonl # 2 records β Personal site
βββ biography.jsonl # 1 record β Complete author profile
βββ agnuxo1-ecosystem-dataset.jsonl # 638 records combined (small records)
βββ agnuxo1-ecosystem-dataset-chunk1.jsonl # 25 records (< 10MB)
βββ agnuxo1-ecosystem-dataset-chunk2.jsonl.gz # Large records (13MB compressed)
βββ agnuxo1-ecosystem-dataset-chunk3.jsonl.gz # Medium records (7MB compressed)
βββ agnuxo1-ecosystem-dataset-chunk4.jsonl.gz # Medium records (8MB compressed)
βββ agnuxo1-ecosystem-dataset-chunk5.jsonl.gz # Medium records (10MB compressed)
βββ agnuxo1-ecosystem-dataset-chunk6.jsonl # 555 records (< 10MB)
βββ README.md # This file
Handling Large Files
Some files exceed Hugging Face's 10MB limit and are stored as gzip compressed chunks with Git LFS:
| Chunk | Original Size | Compressed | Records |
|---|---|---|---|
| CAJAL-chunk2 | 67MB | 13MB | 1 large record |
| CAJAL-chunk3 | 28MB | 7MB | 54 records |
| CAJAL-chunk4 | 28MB | 8MB | 2 records |
| CAJAL-chunk5 | 33MB | 10MB | 1 large record |
| agnuxo1-ecosystem-dataset-chunk2 | 67MB | 13MB | 1 large record |
| agnuxo1-ecosystem-dataset-chunk3 | 28MB | 7MB | 54 records |
| agnuxo1-ecosystem-dataset-chunk4 | 28MB | 8MB | 2 records |
| agnuxo1-ecosystem-dataset-chunk5 | 33MB | 10MB | 1 large record |
To load compressed chunks:
import gzip, json
with gzip.open('CAJAL-chunk2.jsonl.gz', 'rt', encoding='utf-8') as f:
for line in f:
record = json.loads(line)
# Process record
π― Why This Dataset Matters
- Scientific Knowledge Preservation: Complete research trajectory in machine-readable format
- Decentralized Science: Train models on how autonomous peer-review works
- Code + Documentation: Real production code with full context
- Author Attribution: Every record includes ORCID, URLs, and provenance
- Future-Proof: Apache 2.0 license β free for all future LLMs
π Citation
@dataset{angulo2026p2pclaw_ecosystem,
title = {P2PCLAW Ecosystem: Complete Knowledge Base for
Training Scientific AI},
author = {Angulo de Lafuente, Francisco},
year = {2026},
url = {https://huggingface.co/datasets/Agnuxo/p2pclaw-ecosystem-dataset},
license = {Apache-2.0},
orcid = {0009-0001-1634-7063}
}
Built with π₯ by the P2PCLAW Collective β 14 agents, 1 vision, 35 years of research.
"The brain is a world consisting of a number of unexplored continents and great stretches of unknown territory." β Santiago RamΓ³n y Cajal
- Downloads last month
- 87