LongVQ: Long Sequence Modeling with Vector Quantization on Structured Memory

Authors: Zicheng Liu, Li Wang, Siyuan Li, Zedong Wang, Haitao Lin, Stan Z. Li

IJCAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Our experiments on the Long Range Arena benchmark, autoregressive language modeling, and image and speech classification demonstrate the effectiveness of Long VQ.
Researcher Affiliation Academia Zicheng Liu1,2 , Li Wang2 , Siyuan Li1,2 , Zedong Wang2 , Haitao Lin1,2 and Stan Z. Li2 1Zhejiang University, Hangzhou, China 2AI Lab, Research Center for Industries of the Future, Westlake University, Hangzhou, China
Pseudocode No The paper does not contain any pseudocode or algorithm blocks.
Open Source Code No The paper does not provide any explicit statements about open-sourcing code or links to a code repository.
Open Datasets Yes We benchmarked Long VQ on five popular datasets, including Long Range Arena (LRA) [Tay et al., 2020b]... the image dataset s CIFAR [Krizhevsky et al., 2009], the natural language datasets Wiki Text-103 [Merity et al., 2016] and enwik8, and the speech data Speech Command [Warden, 2018].
Dataset Splits Yes The CIFAR-10 dataset s standard train and test split is used, and 10% of the training set is withheld as the validation set.
Hardware Specification Yes All experiments were realized based on NVIDIA A100-80G and Pytorch.
Software Dependencies No The paper mentions 'Pytorch' but does not specify a version number or other software dependencies with versions.
Experiment Setup Yes We used float32 parameters, with bfloat16 precision for most computations. We adopt Adam W as the optimizer with a gradient clip of 0.1. The codebook commit coefficient was always γ = 0.0001, and the codebook EMA rate was always = 0.99. All models were trained with a global batch size of 128 sequences.