Deep Residual Learning in Spiking Neural Networks
Authors: Wei Fang, Zhaofei Yu, Yanqi Chen, Tiejun Huang, Timothée Masquelier, Yonghong Tian
NeurIPS 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We evaluate our SEW Res Net on Image Net, DVS Gesture, and CIFAR10-DVS datasets, and show that SEW Res Net outperforms the state-of-the-art directly trained SNNs in both accuracy and time-steps. |
| Researcher Affiliation | Academia | Wei Fang1,2, Zhaofei Yu1,2 , Yanqi Chen1,2, Tiejun Huang1,2, Timothée Masquelier3, Yonghong Tian1,2 1Department of Computer Science and Technology, Peking University 2Peng Cheng Laboratory, Shenzhen 518055, China 3Centre de Recherche Cerveau et Cognition, UMR5549 CNRS Univ. Toulouse 3 , Toulouse, France |
| Pseudocode | No | No pseudocode or algorithm blocks were found in the paper. |
| Open Source Code | Yes | Our codes are available at https: //github.com/fangwei123456/Spike-Element-Wise-Res Net. |
| Open Datasets | Yes | We evaluate our SEW Res Net on Image Net, DVS Gesture, and CIFAR10-DVS datasets... and the neuromorphic DVS Gesture dataset [1], CIFAR10-DVS dataset [32]. |
| Dataset Splits | Yes | Instead, we use the accuracy on the validation set as the test accuracy, which is the same as [17, 64]. |
| Hardware Specification | No | The paper does not provide specific hardware details (e.g., GPU/CPU models, memory, or cloud instances) used for running the experiments. |
| Software Dependencies | No | The paper mentions software like Pytorch [41] but does not provide specific version numbers for software dependencies or libraries used in their implementation. |
| Experiment Setup | Yes | The IF neuron model is adopted for the static Image Net dataset. ...We design a tiny network named 7B-Net, whose structure is c32k3s1-BN-PLIF-{SEW Block-MPk2s2}*7-FC11. Here c32k3s1 means the convolutional layer with channels 32, kernel size 3, stride 1. MPk2s2 is the max pooling with kernel size 2, stride 2. The symbol {}*7 denotes seven repeated structure, and PLIF denotes the Parametric Leaky-Integrate-and Fire Spiking Neuron... We use the network structure named Wide-7B-Net... The structure of Wide-7B-Net is c64k3s1-BN-PLIF-{SEW Block (c64)-MPk2s2}*4-c128k3s1-BN-PLIF-{SEW Block (c128)-MPk2s2}*3-FC10. |