AutoFormer: Searching Transformers for Visual Recognition

By Minghao Chen, Houwen Peng, Jianlong Fu, Haibin Ling. 2021

350 citations recorded by the source corpus.

Recently, pure transformer-based models have shown great potentials for vision tasks such as image classification and detection. However, the design of transformer networks is challenging. It has been observed that the depth, embedding dimension, and number of heads can largely affect the performance of vision transformers. Previous models configure these dimensions based upon manual crafting. In this work, we propose a new one-shot architecture search framework, namely AutoFormer, dedicated to vision transformer search. AutoFormer entangles the weights of different blocks in the same layers during supernet training. Benefiting from the strategy, the trained supernet allows thousands of subnets to be very well-trained. Specifically, the performance of these subnets with weights inherited from the supernet is comparable to those retrained from scratch. Besides, the searched models, which we refer to AutoFormers, surpass the recent state-of-the-arts such as ViT and DeiT. In particular, AutoFormer-tiny/small/base achieve 74.7%/81.7%/82.4% top-1 accuracy on ImageNet with 5.7M/22.9M/53.7M parameters, respectively. Lastly, we verify the transferability of AutoFormer by providing the performance on downstream benchmarks and distillation experiments. Code and models are available at https://github.com/microsoft/Cream.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

Citations in the corpus

Transformers in Vision: A Survey (2021, background): “[32] propose a one-shot NAS for vision transformers, called AutoFormer.”
BossNAS: Exploring Hybrid CNN-transformers with Block-wisely Self-supervised Neural Architecture Search (2021, background): “Candidates of different channel numbers in our supernet share the weights in a slimmable manner [78, 77, 76, 38, 9].”
SwinFusion: Cross-domain Long-range Learning for General Image Fusion via Swin Transformer (2022, background): “There are many Transformer-based models that have achieved impressive performance in various vision tasks, such as visual recognition [29], [63], [64], object detection [30], [65], [66], tracking [67]–[69], segmentation [31], [70], and image restoration [32], [33], [71].”
TinyViT: Fast Pretraining Distillation for Small Vision Transformers (2022, methodology): “Besides, TinyViT surpasses automatically searched models, such as AutoFormer [11] and GLiT [9].”
Rethinking Vision Transformers for MobileNet Size and Speed (2022, methodology): “Similar to CNNs, architecture search [12, 24, 9, 92, 50, 17, 80], pruning [88], and quantization [54] are also explored for ViTs.”
GLiT: Neural Architecture Search for Global and Local Image Transformer (2021, methodology): “To summarize, the main contributions are as follows: • So far as we know, concurrent with [8], we are the first to explore better transformer architecture by NAS for image classification.”
A Survey on Vision Transformer (2020, methodology): “, Scaling-ViT [269], ViTAS [205], AutoFormer [28] and GLiT [24].”
Parameter-Efficient Fine-Tuning for Large Models: A Comprehensive Survey (2024, methodology): “It utilizes AutoFormer [20], a one-shot NAS algorithm, for the efficient discovery of optimal prompt modules.”
EfficientViT: Memory Efficient Vision Transformer with Cascaded Group Attention (2023, background): “els, the configurations of these components need to be carefully re-designed [7, 8, 39].”
Visformer: The Vision-friendly Transformer (2021, background): “AutoFormer [35] builds a super Transformer network by which they can evaluate the different designs and search efficient Transformer architecture.”
EfficientFormer: Vision Transformers at MobileNet Speed (2022, methodology): “To have a clear understanding of which operations and design choices slow down the inference of ViTs on edge devices, we perform a comprehensive latency analysis over a number of models and operations, as shown in Fig.”
Transformers in Time Series: A Survey (2022, methodology): “GAN GTA [ Chen et al. , 2021c ] combines Transformer with graph-based learning architecture for multivariate time series anomaly detection.”