Picture for Jingwen Leng

Jingwen Leng

Fovea Transformer: Efficient Long-Context Modeling with Structured Fine-to-Coarse Attention

Add code
Nov 13, 2023
Viaarxiv icon

Accelerating Generic Graph Neural Networks via Architecture, Compiler, Partition Method Co-Design

Aug 16, 2023
Figure 1 for Accelerating Generic Graph Neural Networks via Architecture, Compiler, Partition Method Co-Design
Figure 2 for Accelerating Generic Graph Neural Networks via Architecture, Compiler, Partition Method Co-Design
Figure 3 for Accelerating Generic Graph Neural Networks via Architecture, Compiler, Partition Method Co-Design
Figure 4 for Accelerating Generic Graph Neural Networks via Architecture, Compiler, Partition Method Co-Design
Viaarxiv icon

AdaptGear: Accelerating GNN Training via Adaptive Subgraph-Level Kernels on GPUs

May 27, 2023
Figure 1 for AdaptGear: Accelerating GNN Training via Adaptive Subgraph-Level Kernels on GPUs
Figure 2 for AdaptGear: Accelerating GNN Training via Adaptive Subgraph-Level Kernels on GPUs
Figure 3 for AdaptGear: Accelerating GNN Training via Adaptive Subgraph-Level Kernels on GPUs
Figure 4 for AdaptGear: Accelerating GNN Training via Adaptive Subgraph-Level Kernels on GPUs
Viaarxiv icon

Fourier Transformer: Fast Long Range Modeling by Removing Sequence Redundancy with FFT Operator

Add code
May 24, 2023
Figure 1 for Fourier Transformer: Fast Long Range Modeling by Removing Sequence Redundancy with FFT Operator
Figure 2 for Fourier Transformer: Fast Long Range Modeling by Removing Sequence Redundancy with FFT Operator
Figure 3 for Fourier Transformer: Fast Long Range Modeling by Removing Sequence Redundancy with FFT Operator
Figure 4 for Fourier Transformer: Fast Long Range Modeling by Removing Sequence Redundancy with FFT Operator
Viaarxiv icon

Nesting Forward Automatic Differentiation for Memory-Efficient Deep Neural Network Training

Sep 22, 2022
Figure 1 for Nesting Forward Automatic Differentiation for Memory-Efficient Deep Neural Network Training
Figure 2 for Nesting Forward Automatic Differentiation for Memory-Efficient Deep Neural Network Training
Figure 3 for Nesting Forward Automatic Differentiation for Memory-Efficient Deep Neural Network Training
Figure 4 for Nesting Forward Automatic Differentiation for Memory-Efficient Deep Neural Network Training
Viaarxiv icon

ANT: Exploiting Adaptive Numerical Data Type for Low-bit Deep Neural Network Quantization

Add code
Aug 30, 2022
Figure 1 for ANT: Exploiting Adaptive Numerical Data Type for Low-bit Deep Neural Network Quantization
Figure 2 for ANT: Exploiting Adaptive Numerical Data Type for Low-bit Deep Neural Network Quantization
Figure 3 for ANT: Exploiting Adaptive Numerical Data Type for Low-bit Deep Neural Network Quantization
Figure 4 for ANT: Exploiting Adaptive Numerical Data Type for Low-bit Deep Neural Network Quantization
Viaarxiv icon

Efficient Activation Quantization via Adaptive Rounding Border for Post-Training Quantization

Add code
Aug 25, 2022
Figure 1 for Efficient Activation Quantization via Adaptive Rounding Border for Post-Training Quantization
Figure 2 for Efficient Activation Quantization via Adaptive Rounding Border for Post-Training Quantization
Figure 3 for Efficient Activation Quantization via Adaptive Rounding Border for Post-Training Quantization
Figure 4 for Efficient Activation Quantization via Adaptive Rounding Border for Post-Training Quantization
Viaarxiv icon

SALO: An Efficient Spatial Accelerator Enabling Hybrid Sparse Attention Mechanisms for Long Sequences

Jun 29, 2022
Figure 1 for SALO: An Efficient Spatial Accelerator Enabling Hybrid Sparse Attention Mechanisms for Long Sequences
Figure 2 for SALO: An Efficient Spatial Accelerator Enabling Hybrid Sparse Attention Mechanisms for Long Sequences
Figure 3 for SALO: An Efficient Spatial Accelerator Enabling Hybrid Sparse Attention Mechanisms for Long Sequences
Figure 4 for SALO: An Efficient Spatial Accelerator Enabling Hybrid Sparse Attention Mechanisms for Long Sequences
Viaarxiv icon

Transkimmer: Transformer Learns to Layer-wise Skim

Add code
May 15, 2022
Figure 1 for Transkimmer: Transformer Learns to Layer-wise Skim
Figure 2 for Transkimmer: Transformer Learns to Layer-wise Skim
Figure 3 for Transkimmer: Transformer Learns to Layer-wise Skim
Figure 4 for Transkimmer: Transformer Learns to Layer-wise Skim
Viaarxiv icon

SQuant: On-the-Fly Data-Free Quantization via Diagonal Hessian Approximation

Add code
Feb 14, 2022
Figure 1 for SQuant: On-the-Fly Data-Free Quantization via Diagonal Hessian Approximation
Figure 2 for SQuant: On-the-Fly Data-Free Quantization via Diagonal Hessian Approximation
Figure 3 for SQuant: On-the-Fly Data-Free Quantization via Diagonal Hessian Approximation
Figure 4 for SQuant: On-the-Fly Data-Free Quantization via Diagonal Hessian Approximation
Viaarxiv icon