Elias-Abebe-Gasparini/PQID

--- license: cc-by-4.0 language: - en tags: - quantum-computing - qiskit - openqasm - instruction-tuning --- # Parallel Quantum Instruction Dataset (PQID) ## Dataset Description The Parallel Quantum Instruction Dataset (PQID) is a high-fidelity corpus of 10,718 instruction-response pairs designed to bridge the gap between natural language intent and multi-abstraction quantum logic. It is built specifically for instruction-tuning Large Language Models (LLMs) in quantum software engineering. The dataset features a dual-abstraction architecture, pairing user prompts with equivalent quantum circuit generations in both high-level Python (Qiskit) and low-level hardware representation (OPENQASM 3.0). ## Dataset Structure Each row in the dataset represents a single interaction and is formatted as a JSON object: * `instruction`: A natural language request (e.g., "Construct a Bell state circuit"). * `qiskit_code`: The ground-truth Python code utilizing the IBM Qiskit framework. * `qasm_code`: The equivalent ground-truth OPENQASM 3.0 representation. ## Quickstart You can load this dataset directly into your Python environment using the Hugging Face `datasets` library: ```python from datasets import load_dataset dataset = load_dataset("Elias-Abebe-Gasparini/PQID")