There is a newer version of the record available.

Published May 4, 2026 | Version v1.0.0-scientific-data-submission

PQID-Dataset: PQID v1.0.0

  • 1. Yonsei University

Description

PQID v1.0.0 

This record archives the sanitized public GitHub code and documentation snapshot for the Parallel Quantum Instruction Dataset (PQID) v1.0.0.

PQID is a license-aware, quality-audited instruction dataset for quantum programming. It links reconstructed quantum-code source records to role-conditioned instruction rows, paraphrase variants, provenance metadata, validation signals, semantic-consistency metrics, release-governance fields, and public release views.

Dataset scope

The construction-complete PQID object contains 91,719 reconstructed source-corpus records, 91,719 role-conditioned seed instruction rows, 458,595 lineage-preserving paraphrase rows, and 550,314 total construction-complete instruction rows.

The public dataset release is distributed through the Hugging Face dataset repository. The release-ready public views contain 319,782 license-valid rows and 311,724 strict permissive-only public-open rows. Rows associated with explicit no_license status are excluded from public release views and retained only for internal audit.

Archive contents

This Zenodo record should be treated as the reproducibility and documentation package corresponding to the public dataset release. It includes public project documentation, schema and pipeline documentation, release-view summaries, attribution manifests, license-governance summaries, metadata-design evaluation summaries, validation and remediation summaries, semantic-consistency and language-audit summaries, and scripts/notebooks needed to regenerate the public release views and audit summaries.

The machine-learning dataset files themselves are distributed through the Hugging Face dataset repository for convenient downstream access.

Auditability Dashboard

The general dataset construction pipeline and metadata architecture are auditable on the interactive Hugging Face Gradio dashboard on the following address:

Gradio Dashboard 

Reproducibility

The release-export workflow is centered on Stage N of:

PQID/scripts/03_instruction_generation/seed_generation_quality_aware_pipeline.ipynb

Key public scripts and notebooks include:

  • PQID/scripts/export_license_valid_release_views.py
  • PQID/scripts/03_instruction_generation/seed_generation_quality_aware_pipeline.ipynb
  • PQID/scripts/04_metadata_analysis/pqid_metadata_design_and_evaluation.ipynb

Version of record

This DOI identifies the archived public reproducibility package for PQID v1.0.0:

10.5281/zenodo.20024477

Files

PQID-Dataset-v1.0.0-scientific-data-submission.zip

Files (39.1 MB)

Name Size
md5:27cc45c4a4c43f74cbe47ae0c151feb0
39.1 MB Preview Download

Additional details

Software

Repository URL
https://github.com/Elias-Abebe-Gasparini/PQID-Dataset
Programming language
Python , Mermaid
Development Status
Active