Picture for Samuel Felipe dos Santos

Samuel Felipe dos Santos

TransferAttn: Transferable-guided Attention Is All You Need for Video Domain Adaptation

Add code
Jul 01, 2024
Viaarxiv icon

CNNs for JPEGs: A Study in Computational Cost

Add code
Sep 22, 2023
Viaarxiv icon

Budget-Aware Pruning: Handling Multiple Domains with Less Parameters

Add code
Sep 20, 2023
Viaarxiv icon

Parameter Sharing in Budget-Aware Adapters for Multi-Domain Learning

Add code
Oct 14, 2022
Figure 1 for Parameter Sharing in Budget-Aware Adapters for Multi-Domain Learning
Figure 2 for Parameter Sharing in Budget-Aware Adapters for Multi-Domain Learning
Figure 3 for Parameter Sharing in Budget-Aware Adapters for Multi-Domain Learning
Figure 4 for Parameter Sharing in Budget-Aware Adapters for Multi-Domain Learning
Viaarxiv icon

Less is More: Accelerating Faster Neural Networks Straight from JPEG

Add code
Apr 01, 2021
Figure 1 for Less is More: Accelerating Faster Neural Networks Straight from JPEG
Figure 2 for Less is More: Accelerating Faster Neural Networks Straight from JPEG
Figure 3 for Less is More: Accelerating Faster Neural Networks Straight from JPEG
Viaarxiv icon

Deep Learning Towards Edge Computing: Neural Networks Straight from Compressed Data

Add code
Dec 26, 2020
Figure 1 for Deep Learning Towards Edge Computing: Neural Networks Straight from Compressed Data
Figure 2 for Deep Learning Towards Edge Computing: Neural Networks Straight from Compressed Data
Figure 3 for Deep Learning Towards Edge Computing: Neural Networks Straight from Compressed Data
Figure 4 for Deep Learning Towards Edge Computing: Neural Networks Straight from Compressed Data
Viaarxiv icon

Faster and Accurate Compressed Video Action Recognition Straight from the Frequency Domain

Add code
Dec 26, 2020
Figure 1 for Faster and Accurate Compressed Video Action Recognition Straight from the Frequency Domain
Figure 2 for Faster and Accurate Compressed Video Action Recognition Straight from the Frequency Domain
Figure 3 for Faster and Accurate Compressed Video Action Recognition Straight from the Frequency Domain
Figure 4 for Faster and Accurate Compressed Video Action Recognition Straight from the Frequency Domain
Viaarxiv icon