Picture for Zhenglun Kong

Zhenglun Kong

Efficient Pruning of Large Language Model with Adaptive Estimation Fusion

Mar 16, 2024
Figure 1 for Efficient Pruning of Large Language Model with Adaptive Estimation Fusion
Figure 2 for Efficient Pruning of Large Language Model with Adaptive Estimation Fusion
Figure 3 for Efficient Pruning of Large Language Model with Adaptive Estimation Fusion
Figure 4 for Efficient Pruning of Large Language Model with Adaptive Estimation Fusion
Viaarxiv icon

EdgeQAT: Entropy and Distribution Guided Quantization-Aware Training for the Acceleration of Lightweight LLMs on the Edge

Add code
Feb 16, 2024
Viaarxiv icon

Agile-Quant: Activation-Guided Quantization for Faster Inference of LLMs on the Edge

Dec 09, 2023
Figure 1 for Agile-Quant: Activation-Guided Quantization for Faster Inference of LLMs on the Edge
Figure 2 for Agile-Quant: Activation-Guided Quantization for Faster Inference of LLMs on the Edge
Figure 3 for Agile-Quant: Activation-Guided Quantization for Faster Inference of LLMs on the Edge
Figure 4 for Agile-Quant: Activation-Guided Quantization for Faster Inference of LLMs on the Edge
Viaarxiv icon

GPU Accelerated Color Correction and Frame Warping for Real-time Video Stitching

Aug 17, 2023
Figure 1 for GPU Accelerated Color Correction and Frame Warping for Real-time Video Stitching
Figure 2 for GPU Accelerated Color Correction and Frame Warping for Real-time Video Stitching
Figure 3 for GPU Accelerated Color Correction and Frame Warping for Real-time Video Stitching
Figure 4 for GPU Accelerated Color Correction and Frame Warping for Real-time Video Stitching
Viaarxiv icon

You Need Multiple Exiting: Dynamic Early Exiting for Accelerating Unified Vision Language Model

Add code
Nov 21, 2022
Figure 1 for You Need Multiple Exiting: Dynamic Early Exiting for Accelerating Unified Vision Language Model
Figure 2 for You Need Multiple Exiting: Dynamic Early Exiting for Accelerating Unified Vision Language Model
Figure 3 for You Need Multiple Exiting: Dynamic Early Exiting for Accelerating Unified Vision Language Model
Figure 4 for You Need Multiple Exiting: Dynamic Early Exiting for Accelerating Unified Vision Language Model
Viaarxiv icon

Peeling the Onion: Hierarchical Reduction of Data Redundancy for Efficient Vision Transformer Training

Add code
Nov 19, 2022
Figure 1 for Peeling the Onion: Hierarchical Reduction of Data Redundancy for Efficient Vision Transformer Training
Figure 2 for Peeling the Onion: Hierarchical Reduction of Data Redundancy for Efficient Vision Transformer Training
Figure 3 for Peeling the Onion: Hierarchical Reduction of Data Redundancy for Efficient Vision Transformer Training
Figure 4 for Peeling the Onion: Hierarchical Reduction of Data Redundancy for Efficient Vision Transformer Training
Viaarxiv icon

HeatViT: Hardware-Efficient Adaptive Token Pruning for Vision Transformers

Nov 15, 2022
Figure 1 for HeatViT: Hardware-Efficient Adaptive Token Pruning for Vision Transformers
Figure 2 for HeatViT: Hardware-Efficient Adaptive Token Pruning for Vision Transformers
Figure 3 for HeatViT: Hardware-Efficient Adaptive Token Pruning for Vision Transformers
Figure 4 for HeatViT: Hardware-Efficient Adaptive Token Pruning for Vision Transformers
Viaarxiv icon

The Lottery Ticket Hypothesis for Vision Transformers

Add code
Nov 02, 2022
Figure 1 for The Lottery Ticket Hypothesis for Vision Transformers
Figure 2 for The Lottery Ticket Hypothesis for Vision Transformers
Figure 3 for The Lottery Ticket Hypothesis for Vision Transformers
Figure 4 for The Lottery Ticket Hypothesis for Vision Transformers
Viaarxiv icon

Layer Freezing & Data Sieving: Missing Pieces of a Generic Framework for Sparse Training

Add code
Sep 22, 2022
Figure 1 for Layer Freezing & Data Sieving: Missing Pieces of a Generic Framework for Sparse Training
Figure 2 for Layer Freezing & Data Sieving: Missing Pieces of a Generic Framework for Sparse Training
Figure 3 for Layer Freezing & Data Sieving: Missing Pieces of a Generic Framework for Sparse Training
Figure 4 for Layer Freezing & Data Sieving: Missing Pieces of a Generic Framework for Sparse Training
Viaarxiv icon

SPViT: Enabling Faster Vision Transformers via Soft Token Pruning

Add code
Dec 27, 2021
Figure 1 for SPViT: Enabling Faster Vision Transformers via Soft Token Pruning
Figure 2 for SPViT: Enabling Faster Vision Transformers via Soft Token Pruning
Figure 3 for SPViT: Enabling Faster Vision Transformers via Soft Token Pruning
Figure 4 for SPViT: Enabling Faster Vision Transformers via Soft Token Pruning
Viaarxiv icon