Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1].
Spike2Former: Efficient Spiking Transformer for High-performance Image Segmentation
Authors: Zhenxin Lei, Man Yao, Jiakui Hu, Xinhao Luo, Yanye Lu, Bo Xu, Guoqi Li
AAAI 2025 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We set a new state-of-the-art for SNNs in various semantic segmentation datasets, with a significant improvement of +12.7% m Io U and 5.0 efficiency on ADE20K, +14.3% m Io U and 5.2 efficiency on VOC2012, and +9.1% m Io U and 6.6 efficiency on City Scapes. |
| Researcher Affiliation | Academia | 1University of Chinese Academy of Sciences 2Institute of Automation, Key Laboratory of Brain Cognition and Brain-inspired Intelligence Technology, Chinese Academy of Sciences 3Institute of Medical Technology, Peking University Health Science Center, Peking University 4National Biomedical Imaging Center, Peking University |
| Pseudocode | No | The paper describes methods through mathematical formulations and architectural diagrams (e.g., Figure 1, Equations 1-21) but does not include a dedicated pseudocode or algorithm block. |
| Open Source Code | No | The paper does not contain any explicit statement about releasing code, nor does it provide a link to a code repository. |
| Open Datasets | Yes | We conduct semantic segmentation on ADE20k (Zhou et al. 2017), City Scapes (Cordts et al. 2016), and Pascal VOC2012 (Everingham et al. 2010) datasets. |
| Dataset Splits | No | The paper mentions input sizes for datasets (e.g., 'ADE20k Input Size 512 512') and discusses training settings (e.g., 'Learning Rate', 'Optimizer', 'Training Steps'), but does not explicitly provide details about training, validation, or test dataset splits in the main text. It states 'More training details can be found in the Appendix.', but the appendix is not provided. |
| Hardware Specification | No | The paper does not provide specific hardware details (e.g., GPU models, CPU types, or memory specifications) used for running its experiments. |
| Software Dependencies | No | The paper mentions using 'Adam W' as an optimizer and 'Meta Spikeformer' as a backbone, but does not provide specific version numbers for any programming languages, libraries, or software frameworks used in the implementation or experimentation. |
| Experiment Setup | Yes | Table 2: Hyper-parameters setting in Spike2Former. Input Size 512 512 (ADE20k), 512 1024 (City Scapes), 512 512 (Pascal VOC2012). Learning Rate 2e-4 (ADE20k), 2e-3 (City Scapes), 2e-3 (Pascal VOC2012). Optimizer Adam W (all). Training Steps 160k (ADE20k), 90k (City Scapes), 80k (Pascal VOC2012). |