Picture for Zhenyi Lu

Zhenyi Lu

Extrapolating and Decoupling Image-to-Video Generation Models: Motion Modeling is Easier Than You Think

Add code
Mar 02, 2025
Viaarxiv icon

Make LoRA Great Again: Boosting LoRA with Adaptive Singular Values and Mixture-of-Experts Optimization Alignment

Add code
Feb 26, 2025
Viaarxiv icon

Twin-Merging: Dynamic Integration of Modular Expertise in Model Merging

Add code
Jun 17, 2024
Figure 1 for Twin-Merging: Dynamic Integration of Modular Expertise in Model Merging
Figure 2 for Twin-Merging: Dynamic Integration of Modular Expertise in Model Merging
Figure 3 for Twin-Merging: Dynamic Integration of Modular Expertise in Model Merging
Figure 4 for Twin-Merging: Dynamic Integration of Modular Expertise in Model Merging
Viaarxiv icon

On Giant's Shoulders: Effortless Weak to Strong by Dynamic Logits Fusion

Add code
Jun 17, 2024
Figure 1 for On Giant's Shoulders: Effortless Weak to Strong by Dynamic Logits Fusion
Figure 2 for On Giant's Shoulders: Effortless Weak to Strong by Dynamic Logits Fusion
Figure 3 for On Giant's Shoulders: Effortless Weak to Strong by Dynamic Logits Fusion
Figure 4 for On Giant's Shoulders: Effortless Weak to Strong by Dynamic Logits Fusion
Viaarxiv icon

Mitigating Boundary Ambiguity and Inherent Bias for Text Classification in the Era of Large Language Models

Add code
Jun 11, 2024
Figure 1 for Mitigating Boundary Ambiguity and Inherent Bias for Text Classification in the Era of Large Language Models
Figure 2 for Mitigating Boundary Ambiguity and Inherent Bias for Text Classification in the Era of Large Language Models
Figure 3 for Mitigating Boundary Ambiguity and Inherent Bias for Text Classification in the Era of Large Language Models
Figure 4 for Mitigating Boundary Ambiguity and Inherent Bias for Text Classification in the Era of Large Language Models
Viaarxiv icon

Improving Low-resource Prompt-based Relation Representation with Multi-view Decoupling Learning

Add code
Jan 15, 2024
Figure 1 for Improving Low-resource Prompt-based Relation Representation with Multi-view Decoupling Learning
Figure 2 for Improving Low-resource Prompt-based Relation Representation with Multi-view Decoupling Learning
Figure 3 for Improving Low-resource Prompt-based Relation Representation with Multi-view Decoupling Learning
Figure 4 for Improving Low-resource Prompt-based Relation Representation with Multi-view Decoupling Learning
Viaarxiv icon

MIRACLE: Towards Personalized Dialogue Generation with Latent-Space Multiple Personal Attribute Control

Add code
Oct 22, 2023
Viaarxiv icon