Data augmentations are important in training high-performance 3D object
...
3D object detection in point clouds is a core component for modern robot...
Developing neural models that accurately understand objects in 3D point
...
Cross-entropy loss and focal loss are the most common choices when train...
It is commonly believed that high internal resolution combined with expe...
We propose Occupancy Flow Fields, a new representation for motion foreca...
We present a combined scaling method called BASIC that achieves 85.7
zer...
Transformers have attracted increasing interests in computer vision, but...
This paper introduces EfficientNetV2, a new family of convolutional netw...
Data augmentation has become a de facto component for training
high-perf...
We present Mobile Video Networks (MoViNets), a family of computation and...
Neural architectures and hardware accelerators have been two driving for...
Neural Architecture Search (NAS), together with model scaling, has shown...
Transformers have emerged as a powerful tool for a broad range of natura...
Neural networks are sensitive to hyper-parameter and architecture choice...
EfficientNets are a family of state-of-the-art image classification mode...
Recently, SpineNet has demonstrated promising results on object detectio...
Shape and texture are two prominent and complementary cues for recognizi...
It is commonly believed that networks cannot be both accurate and robust...
Neural Architecture Search (NAS) has achieved significant progress in pu...
Ensembling is a simple and popular technique for boosting evaluation
per...
Inverted bottleneck layers, which are built upon depthwise convolutions,...
Neural architecture search (NAS) has shown promising results discovering...
Convolutional neural networks typically encode an input image into a ser...
Adversarial examples are commonly viewed as a threat to ConvNets. Here w...
Standard Knowledge Distillation (KD) approaches distill the knowledge of...
Model efficiency has become increasingly important in computer vision. I...
Depthwise convolution is becoming increasingly popular in modern efficie...
Learning to represent videos is a very challenging task both algorithmic...
Convolutional Neural Networks (ConvNets) are commonly developed at a fix...
We present the next generation of MobileNets based on a combination of
c...
Designing convolutional neural networks (CNN) models for mobile devices ...