Picture for Chunhui Zhang

Chunhui Zhang

Scaled Supervision is an Implicit Lipschitz Regularizer

Add code
Mar 19, 2025
Viaarxiv icon

Superficial Self-Improved Reasoners Benefit from Model Merging

Add code
Mar 03, 2025
Viaarxiv icon

Modality-Aware Neuron Pruning for Unlearning in Multimodal Large Language Models

Add code
Feb 21, 2025
Viaarxiv icon

Pretrained Image-Text Models are Secretly Video Captioners

Add code
Feb 19, 2025
Viaarxiv icon

Learning Musical Representations for Music Performance Question Answering

Add code
Feb 10, 2025
Viaarxiv icon

Temporal Working Memory: Query-Guided Segment Refinement for Enhanced Multimodal Understanding

Add code
Feb 09, 2025
Figure 1 for Temporal Working Memory: Query-Guided Segment Refinement for Enhanced Multimodal Understanding
Figure 2 for Temporal Working Memory: Query-Guided Segment Refinement for Enhanced Multimodal Understanding
Figure 3 for Temporal Working Memory: Query-Guided Segment Refinement for Enhanced Multimodal Understanding
Figure 4 for Temporal Working Memory: Query-Guided Segment Refinement for Enhanced Multimodal Understanding
Viaarxiv icon

Is It Navajo? Accurate Language Detection in Endangered Athabaskan Languages

Add code
Jan 27, 2025
Viaarxiv icon

MambaTrack: Exploiting Dual-Enhancement for Night UAV Tracking

Add code
Nov 24, 2024
Figure 1 for MambaTrack: Exploiting Dual-Enhancement for Night UAV Tracking
Figure 2 for MambaTrack: Exploiting Dual-Enhancement for Night UAV Tracking
Figure 3 for MambaTrack: Exploiting Dual-Enhancement for Night UAV Tracking
Figure 4 for MambaTrack: Exploiting Dual-Enhancement for Night UAV Tracking
Viaarxiv icon

Towards Underwater Camouflaged Object Tracking: An Experimental Evaluation of SAM and SAM 2

Add code
Sep 25, 2024
Viaarxiv icon

Distilling Channels for Efficient Deep Tracking

Add code
Sep 18, 2024
Figure 1 for Distilling Channels for Efficient Deep Tracking
Figure 2 for Distilling Channels for Efficient Deep Tracking
Figure 3 for Distilling Channels for Efficient Deep Tracking
Figure 4 for Distilling Channels for Efficient Deep Tracking
Viaarxiv icon