Sequencer: Deep LSTM for Image Classification

Authors: Yuki Tatsunami, Masato Taki

NeurIPS 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental In this section, we compare Sequencers with previous studies on the IN-1K benchmark [39]. We also carry out ablation studies, transfer learning studies, and analysis of the results to demonstrate the effectiveness of Sequencers.
Researcher Affiliation Collaboration Yuki Tatsunami1,2 Masato Taki1 1Rikkyo University, Tokyo, Japan 2Any Tech Co., Ltd., Tokyo, Japan
Pseudocode Yes The Py Torch-like pseudocode is shown in Appendix B.1.
Open Source Code Yes Our source code is available at https://github.com/okojoalg/sequencer.
Open Datasets Yes We utilize IN-1K [39], which has 1000 classes and contains 1,281,167 training images and 50,000 validation images.
Dataset Splits Yes We utilize IN-1K [39], which has 1000 classes and contains 1,281,167 training images and 50,000 validation images.
Hardware Specification Yes Training and inference throughput and their peak memory were measured with 16 images per batch on a single V100 GPU.
Software Dependencies No The paper mentions using 'Py Torch [56] and timm [80] library' but does not specify the version numbers for these software dependencies, which is required for reproducibility.
Experiment Setup Yes We adopt Adam W optimizer [50]. Following the previous study [72], we adopt the base learning rate batch size 512 5 10^4. The batch sizes for Sequencer2D-S, Sequencer2D-M, and Sequencer2D-L are 2048, 1536, and 1024, respectively. As a regularization method, stochastic depth [30] and label smoothing [66] are employed. As data augmentation methods, mixup [87], cutout [12], cutmix [85], random erasing [88], and randaugment [11] are applied.