SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2403.16552 · NeurIPS · 2024

QKFormer: Hierarchical Spiking Transformer using Q-K Attention

Han Zhang, Yonghong Tian, Li Yuan, Zhengyu Ma, Zhaokun Zhou, Chenlin Zhou, Liutao Yu, Liwei Huang, Xiaopeng Fan, Huihui Zhou

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 5 functions out of this paper's own repositories and ran 3 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.

RepositoryRoleRan
zhouchenlin2096/QKFormer canonical 2 of 2
Fancyssc/Spiking-Transformers pwc_unofficial 1 of 3
FunctionStatusWhere it lives
accuracy Ran zhouchenlin2096/QKFormer/cifar10-dvs/utils.py
pointer only (licence: NONE) · get_code("4134b8420b4c48b5")
build_transform Ran Fancyssc/Spiking-Transformers/datasets.py
code served (permissive licence) · get_code("aeded02f9fc26708")
unpack_len1_tuple Ran zhouchenlin2096/QKFormer/cifar10-dvs/monitor.py
pointer only (licence: NONE) · get_code("68edbf1d7855d1bc")
build_dataset Not yet run Fancyssc/Spiking-Transformers/datasets.py
code served (permissive licence) · get_code("99c6a11c5039db76")
unpack_mix_param Not yet run Fancyssc/Spiking-Transformers/datasets.py
code served (permissive licence) · get_code("137a4ac791af3cd7")

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Spiking Transformers, which integrate Spiking Neural Networks (SNNs) with Transformer architectures, have attracted significant attention due to their potential for low energy consumption and high performance. However, there remains a substantial gap in performance between SNNs and Artificial Neural Networks (ANNs). To narrow this gap, we have developed QKFormer, a direct training spiking transformer with the following features: i) Linear complexity and high energy efficiency, the novel spike-form Q-K attention module efficiently models the token or channel attention through binary vectors and enables the construction of larger models. ii) Multi-scale spiking representation, achieved by a hierarchical structure with the different number of tokens across blocks. iii) Spiking Patch Embedding with Deformed Shortcut (SPEDS), enhances spiking information transmission and integration, thus improving overall performance. It is shown that QKFormer achieves significantly superior performance over existing state-of-the-art SNN models on various mainstream datasets. Notably, with comparable size to Spikformer (66.34 M, 74.81%), QKFormer (64.96 M) achieves a groundbreaking top-1 accuracy of 85.65% on ImageNet-1k, substantially outperforming Spikformer by 10.84%. To our best knowledge, this is the first time that directly training SNNs have exceeded 85% accuracy on ImageNet-1K. The code and models are available at https://github.com/zhouchenlin2096/QKFormer.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2403.16552")
get_code_for_paper("2403.16552")
have("2403.16552")

Connect an agent — have() is free.