Dataset Viewer

The dataset viewer is not available because its heuristics could not detect any supported data files. You can try uploading some data files, or configuring the data files location manually.

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

Dvidata

A collection of original AI training datasets for the Clobax ecosystem — a family of open computer architecture projects by Avyaan Mishra.

This repository contains documentation-style instruction/response pairs covering the Doublers, Doton, Doton-OS, DOB, DULO, Dormals, and related open specifications.

Contents

File Entries Size Description
dvidvar_doublers_10000.json 10,000 590 MB Doublers open computer architecture (ISA + Command Architecture)
all_datasets.jsonl 270,261 518 MB Combined dataset (JSONL format)
expanded_project_dataset.json 18,359 8.1 MB Expanded multi-project dataset
dob-vcs.json 6,333 2.9 MB DOB — Open Version Control System specification
dulo.json 5,962 2.6 MB DULO — Open Spreadsheet Engine specification
doton.json 5,399 2.4 MB Doton — Microkernel specification
dormals.json 353 0.2 MB Dormals — AI companion knowledge base
doton-os.json 312 0.1 MB Doton-OS — Open operating system ecosystem
metadata.json Dataset metadata
LICENSE.txt GPL-3.0 license

Split Datasets

Large files are split into parts, each strictly under 25 MB, for easier storage and upload.

ALL DATASET FOLDER/

20 parts: all_datasets_part001.jsonlall_datasets_part020.jsonl

Dvidvar_Doublers 10000/

23 parts: dvidvar_doublers_10000_part001.jsondvidvar_doublers_10000_part023.json

Each JSON part is a valid standalone array. Each JSONL part is a valid line-delimited JSON file.

Data Format

Each entry follows a consistent documentation-Q&A structure:

{
  "id": "DVD00001",
  "version": "1.0",
  "source": "Doublers/Dvidvar documentation",
  "category": "documentation",
  "language": "English",
  "instruction": "Explain the section about # Doublers in the Doublers or Dvidvar documentation.",
  "input": "",
  "response": "# Doublers\n\n> **An open computer architecture for anyone.**...",
  "tags": ["doublers", "documentation", "specification"]
}

Fields

Field Description
id Unique identifier (e.g. DVD00001, DOT00001, DUL00001)
version Dataset version
source Origin of the documentation
category Type of content (e.g. documentation)
language Language of the entry
instruction The prompt / question
input Optional additional input (may be empty)
response The answer / explanation
tags Optional topic tags

Projects Covered

  • Doublers — Open computer architecture (ISA + Command Architecture)
  • Doton — Open microkernel specification
  • Doton-OS — Open operating system ecosystem
  • DOB (Dob-VCS) — Open version control system specification
  • DULO — Open spreadsheet engine specification
  • Dormals — AI companion knowledge base
  • LinkLoader, Deshell, Deson — Additional Clobax ecosystem projects

Usage

Load a JSON dataset

import json

with open("dvidvar_doublers_10000.json", "r", encoding="utf-8") as f:
    data = json.load(f)
print(f"Loaded {len(data)} entries")

Load all split JSON parts

import json

data = []
for i in range(1, 24):
    with open(f"Dvidvar_Doublers 10000/dvidvar_doublers_10000_part{i:03d}.json", "r", encoding="utf-8") as f:
        data.extend(json.load(f))
print(f"Loaded {len(data)} entries")

Read a JSONL dataset

import json

with open("all_datasets.jsonl", "r", encoding="utf-8") as f:
    for line in f:
        if line.strip():
            entry = json.loads(line)
            # process entry

License

GPL-3.0-or-later — Copyright (C) 2026 Avyaan Mishra.

This repository includes original Clobax source code, model architecture, utilities, training scripts, tokenizer tools, and original datasets (Dvidvar, Doton, Doton-OS, DOB, DULO, Dormals, LinkLoader, Deshell, Deson, and other original Clobax ecosystem projects).

Note on third-party datasets: This repository may reference or provide scripts to obtain third-party datasets (CodeSearchNet, CodeXGLUE, HumanEval, MBPP, DS-1000, SWE-bench Lite). even though the folders/files of those third party datasets are not physically included, the all dataset jsonl files are having their knowledge hardcoded in itself, making this Dataset suitable for training a independent model . These are not covered by the GPL license and remain the property of their respective authors. Users are responsible for complying with their original license terms.

See LICENSE.txt for full details.

also, I have many project on Github

visit the account "Pandeykhushbu29-eng" to know more

Also, i will be too busy to join or correct things. but i will TRY MY BEST to come back, even for a few minutes, whenever possible!!!

So due to this, this Project belongs to YOU ALL! i am just the creator, not the controller of this project. This project belongs to all of you!

Also, check out my all my projects

[https://github.com/pandeykhushbu29-eng/Dvidvar]

[https://github.com/pandeykhushbu29-eng/Dormals-AI-companion]

[https://github.com/pandeykhushbu29-eng/Dormals]

[https://github.com/pandeykhushbu29-eng/DOARCE]

[https://github.com/pandeykhushbu29-eng/Dob-VCS]

[https://github.com/pandeykhushbu29-eng/Doton-OS]

[https://github.com/pandeykhushbu29-eng/Doton]

[https://github.com/pandeykhushbu29-eng/DULO]

[https://huggingface.co/spaces/Clobax-Creator/Dvi-Coder.web]

(c) Avyaan Mishra / Clobax creator / Doub.creator_00001(pandeykhushbu29-eng) --(all these are accounts of where my projects are)
(c) Dvidata Community (People who have contributed or even used this project or its parts for anything are in community)

Downloads last month
107