Zobrazeno 1 - 10
of 16
pro vyhledávání: '"Shen, Maying"'
In recent years, the data collected for artificial intelligence has grown to an unmanageable amount. Particularly within industrial applications, such as autonomous vehicles, model training computation budgets are being exceeded while model performan
Externí odkaz:
http://arxiv.org/abs/2409.13860
The cornerstone of autonomous vehicles (AV) is a solid perception system, where camera encoders play a crucial role. Existing works usually leverage pre-trained Convolutional Neural Networks (CNN) or Vision Transformers (ViTs) designed for general vi
Externí odkaz:
http://arxiv.org/abs/2407.07276
As we push the boundaries of performance in various vision tasks, the models grow in size correspondingly. To keep up with this growth, we need very aggressive pruning techniques for efficient inference and deployment on edge devices. Existing prunin
Externí odkaz:
http://arxiv.org/abs/2406.12079
Data often arrives in sequence over time in real-world deep learning applications such as autonomous driving. When new training data is available, training the model from scratch undermines the benefit of leveraging the learned knowledge, leading to
Externí odkaz:
http://arxiv.org/abs/2406.04484
Robustness and compactness are two essential attributes of deep learning models that are deployed in the real world. The goals of robustness and compactness may seem to be at odds, since robustness requires generalization across domains, while the pr
Externí odkaz:
http://arxiv.org/abs/2306.14306
Structured channel pruning has been shown to significantly accelerate inference time for convolution neural networks (CNNs) on modern hardware, with a relatively minor loss of network accuracy. Recent works permanently zero these channels during trai
Externí odkaz:
http://arxiv.org/abs/2211.02206
Structural pruning can simplify network architecture and improve inference speed. We propose Hardware-Aware Latency Pruning (HALP) that formulates structural pruning as a global resource allocation optimization problem, aiming at maximizing the accur
Externí odkaz:
http://arxiv.org/abs/2210.06659
Pruning enables appealing reductions in network memory footprint and time complexity. Conventional post-training pruning techniques lean towards efficient inference while overlooking the heavy computation for training. Recent exploration of pre-train
Externí odkaz:
http://arxiv.org/abs/2110.12007
Structural pruning can simplify network architecture and improve inference speed. We propose Hardware-Aware Latency Pruning (HALP) that formulates structural pruning as a global resource allocation optimization problem, aiming at maximizing the accur
Externí odkaz:
http://arxiv.org/abs/2110.10811
Transformers yield state-of-the-art results across many tasks. However, their heuristically designed architecture impose huge computational costs during inference. This work aims on challenging the common design philosophy of the Vision Transformer (
Externí odkaz:
http://arxiv.org/abs/2110.04869