Picture for Christos-Savvas Bouganis

Christos-Savvas Bouganis

SMOF: Streaming Modern CNNs on FPGAs with Smart Off-Chip Eviction

Add code
Mar 27, 2024
Figure 1 for SMOF: Streaming Modern CNNs on FPGAs with Smart Off-Chip Eviction
Figure 2 for SMOF: Streaming Modern CNNs on FPGAs with Smart Off-Chip Eviction
Figure 3 for SMOF: Streaming Modern CNNs on FPGAs with Smart Off-Chip Eviction
Figure 4 for SMOF: Streaming Modern CNNs on FPGAs with Smart Off-Chip Eviction
Viaarxiv icon

Understanding Why Label Smoothing Degrades Selective Classification and How to Fix It

Add code
Mar 19, 2024
Viaarxiv icon

SATAY: A Streaming Architecture Toolflow for Accelerating YOLO Models on FPGA Devices

Add code
Sep 04, 2023
Figure 1 for SATAY: A Streaming Architecture Toolflow for Accelerating YOLO Models on FPGA Devices
Figure 2 for SATAY: A Streaming Architecture Toolflow for Accelerating YOLO Models on FPGA Devices
Figure 3 for SATAY: A Streaming Architecture Toolflow for Accelerating YOLO Models on FPGA Devices
Figure 4 for SATAY: A Streaming Architecture Toolflow for Accelerating YOLO Models on FPGA Devices
Viaarxiv icon

Mixed-TD: Efficient Neural Network Accelerator with Layer-Specific Tensor Decomposition

Add code
Jun 22, 2023
Figure 1 for Mixed-TD: Efficient Neural Network Accelerator with Layer-Specific Tensor Decomposition
Figure 2 for Mixed-TD: Efficient Neural Network Accelerator with Layer-Specific Tensor Decomposition
Figure 3 for Mixed-TD: Efficient Neural Network Accelerator with Layer-Specific Tensor Decomposition
Figure 4 for Mixed-TD: Efficient Neural Network Accelerator with Layer-Specific Tensor Decomposition
Viaarxiv icon

fpgaHART: A toolflow for throughput-oriented acceleration of 3D CNNs for HAR onto FPGAs

Add code
May 31, 2023
Figure 1 for fpgaHART: A toolflow for throughput-oriented acceleration of 3D CNNs for HAR onto FPGAs
Figure 2 for fpgaHART: A toolflow for throughput-oriented acceleration of 3D CNNs for HAR onto FPGAs
Figure 3 for fpgaHART: A toolflow for throughput-oriented acceleration of 3D CNNs for HAR onto FPGAs
Figure 4 for fpgaHART: A toolflow for throughput-oriented acceleration of 3D CNNs for HAR onto FPGAs
Viaarxiv icon

FMM-X3D: FPGA-based modeling and mapping of X3D for Human Action Recognition

Add code
May 29, 2023
Figure 1 for FMM-X3D: FPGA-based modeling and mapping of X3D for Human Action Recognition
Figure 2 for FMM-X3D: FPGA-based modeling and mapping of X3D for Human Action Recognition
Figure 3 for FMM-X3D: FPGA-based modeling and mapping of X3D for Human Action Recognition
Figure 4 for FMM-X3D: FPGA-based modeling and mapping of X3D for Human Action Recognition
Viaarxiv icon

ATHEENA: A Toolflow for Hardware Early-Exit Network Automation

Add code
Apr 17, 2023
Figure 1 for ATHEENA: A Toolflow for Hardware Early-Exit Network Automation
Figure 2 for ATHEENA: A Toolflow for Hardware Early-Exit Network Automation
Figure 3 for ATHEENA: A Toolflow for Hardware Early-Exit Network Automation
Figure 4 for ATHEENA: A Toolflow for Hardware Early-Exit Network Automation
Viaarxiv icon

HARFLOW3D: A Latency-Oriented 3D-CNN Accelerator Toolflow for HAR on FPGA Devices

Add code
Apr 11, 2023
Figure 1 for HARFLOW3D: A Latency-Oriented 3D-CNN Accelerator Toolflow for HAR on FPGA Devices
Figure 2 for HARFLOW3D: A Latency-Oriented 3D-CNN Accelerator Toolflow for HAR on FPGA Devices
Figure 3 for HARFLOW3D: A Latency-Oriented 3D-CNN Accelerator Toolflow for HAR on FPGA Devices
Figure 4 for HARFLOW3D: A Latency-Oriented 3D-CNN Accelerator Toolflow for HAR on FPGA Devices
Viaarxiv icon

Window-Based Early-Exit Cascades for Uncertainty Estimation: When Deep Ensembles are More Efficient than Single Models

Add code
Mar 14, 2023
Figure 1 for Window-Based Early-Exit Cascades for Uncertainty Estimation: When Deep Ensembles are More Efficient than Single Models
Figure 2 for Window-Based Early-Exit Cascades for Uncertainty Estimation: When Deep Ensembles are More Efficient than Single Models
Figure 3 for Window-Based Early-Exit Cascades for Uncertainty Estimation: When Deep Ensembles are More Efficient than Single Models
Figure 4 for Window-Based Early-Exit Cascades for Uncertainty Estimation: When Deep Ensembles are More Efficient than Single Models
Viaarxiv icon

SVD-NAS: Coupling Low-Rank Approximation and Neural Architecture Search

Add code
Aug 22, 2022
Figure 1 for SVD-NAS: Coupling Low-Rank Approximation and Neural Architecture Search
Figure 2 for SVD-NAS: Coupling Low-Rank Approximation and Neural Architecture Search
Figure 3 for SVD-NAS: Coupling Low-Rank Approximation and Neural Architecture Search
Figure 4 for SVD-NAS: Coupling Low-Rank Approximation and Neural Architecture Search
Viaarxiv icon