Q-La-LoRA: Quantized Layer-wise Adaptive Low-Rank Adaptation for Memory-Efficient LLM Fine-Tuning

Published: 15 April 2026| Version 1 | DOI: 10.17632/mckxyfjpv8.1
Contributor:
Arun Balaji V

Description

Code and experimental results supporting the paper "Q-La-LoRA". Includes training scripts, hyperparameter configurations, and benchmark accuracy results on 8 commonsense reasoning tasks using LLaMA-2-7B.

Files

Institutions

Categories

Artificial Intelligence, Natural Language Processing, Machine Learning

Funders

Licence