Description
MetaFormer is a general architecture abstracted from Transformers by not specifying the token mixer.
Papers Using This Method
EMF: Event Meta Formers for Event-based Real-time Traffic Object Detection2025-04-05Foundation Models For Seismic Data Processing: An Extensive Review2025-03-31SDTrack: A Baseline for Event-based Tracking via Spiking Neural Networks2025-03-09Static Key Attention in Vision2024-12-09Aberration Correcting Vision Transformers for High-Fidelity Metalens Imaging2024-12-05MVFormer: Diversifying Feature Normalization and Token Mixing for Efficient Vision Transformers2024-11-28MetaFormer and CNN Hybrid Model for Polyp Image Segmentation2024-09-16MetaSeg: MetaFormer-based Global Contexts-aware Network for Efficient Semantic Segmentation2024-08-14ParFormer: A Vision Transformer with Parallel Mixer and Sparse Channel Attention Patch Embedding2024-03-22HyenaPixel: Global Image Context with Convolutions2024-02-29Decision ConvFormer: Local Filtering in MetaFormer is Sufficient for Decision Making2023-10-04SPANet: Frequency-balancing Token Mixer using Spectral Pooling Aggregation Modulation2023-08-22Pretraining is All You Need: A Multi-Atlas Enhanced Transformer Framework for Autism Spectrum Disorder Classification2023-07-04M^2UNet: MetaFormer Multi-scale Upsampling Network for Polyp Segmentation2023-06-14A Mask Free Neural Network for Monaural Speech Enhancement2023-06-07DeformableFormer: Classification of Endoscopic Ultrasound Guided Fine Needle Biopsy in Pancreatic Diseases2023-04-21FFT-based Dynamic Token Mixer for Vision2023-03-07MetaFormer Baselines for Vision2022-10-24MetaFormer: A Unified Meta Framework for Fine-Grained Recognition2022-03-05MetaFormer Is Actually What You Need for Vision2021-11-22