Log In Sign Up

LambdaNetworks: Modeling Long-Range Interactions Without Attention

by   Irwan Bello, et al.

We present lambda layers – an alternative framework to self-attention – for capturing long-range interactions between an input and structured contextual information (e.g. a pixel surrounded by other pixels). Lambda layers capture such interactions by transforming available contexts into linear functions, termed lambdas, and applying these linear functions to each input separately. Similar to linear attention, lambda layers bypass expensive attention maps, but in contrast, they model both content and position-based interactions which enables their application to large structured inputs such as images. The resulting neural network architectures, LambdaNetworks, significantly outperform their convolutional and attentional counterparts on ImageNet classification, COCO object detection and COCO instance segmentation, while being more computationally efficient. Additionally, we design LambdaResNets, a family of hybrid architectures across different scales, that considerably improves the speed-accuracy tradeoff of image classification models. LambdaResNets reach excellent accuracies on ImageNet while being 3.2 - 4.4x faster than the popular EfficientNets on modern machine learning accelerators. When training with an additional 130M pseudo-labeled images, LambdaResNets achieve up to a 9.5x speed-up over the corresponding EfficientNet checkpoints.


Attention Augmented Convolutional Networks

Convolutional networks have been the paradigm of choice in many computer...

Scaling Local Self-Attention For Parameter Efficient Visual Backbones

Self-attention has the promise of improving computer vision systems due ...

XCiT: Cross-Covariance Image Transformers

Following their success in natural language processing, transformers hav...

Rethinking Query-Key Pairwise Interactions in Vision Transformers

Vision Transformers have achieved state-of-the-art performance in many v...

Attention on Classification for Fire Segmentation

Detection and localization of fire in images and videos are important in...

MorphMLP: A Self-Attention Free, MLP-Like Backbone for Image and Video

Self-attention has become an integral component of the recent network ar...

Code Repositories


Implementing Lambda Networks using Pytorch

view repo


Lambda Networks implemented in PyTorch

view repo


Implementation of LambdaNetworks, a framework for capturing long-range interaction between structured contextual information. Tensorflow-2.x implementation.

view repo


TensorFlow2 Implementation of LambdaNetworks

view repo