"QCHFT: Quantum Cross-Hybrid Fine-Tuning for LLMs"
收藏DataCite Commons2026-05-01 更新2026-05-03 收录
下载链接：
https://ieee-dataport.org/documents/qchft-quantum-cross-hybrid-fine-tuning-llms-0
下载链接
链接失效反馈官方服务：
资源简介：
"This repository contains the datasets and experimental results associated with the paper \"QCHFT: Quantum Cross-Hybrid Fine-Tuning for LLMs\". The paper proposes a suite of parameter-efficient fine-tuning (PEFT) methods using Parameterized Quantum Circuits (PQCs) to adapt large language models (LLMs). The provided data includes a primary text-conditioned regression dataset for multi-tier IoT resource allocation, alongside two downsampled natural language processing benchmarks (SST-2 and HellaSwag). The results encompass the fine-tuning performance of five causal LLM backbones\u2014scaling from 125M to 2B parameters (Mamba2, GPT-Neo, Pythia, Llama-3.2, and Qwen-3.5) \u2014using classical baselines (e.g., LoRA) and the proposed QCHFT adapters (MPS, APC, HPC, and AQB)."
提供机构：
IEEE DataPort
创建时间：
2026-05-01
