Picture for Xiaoye Qu

Xiaoye Qu

SEE: Continual Fine-tuning with Sequential Ensemble of Experts

Add code
Apr 09, 2025
Viaarxiv icon

A Survey of Efficient Reasoning for Large Reasoning Models: Language, Multimodality, and Beyond

Add code
Mar 27, 2025
Viaarxiv icon

From Head to Tail: Towards Balanced Representation in Large Vision-Language Models through Adaptive Data Calibration

Add code
Mar 17, 2025
Viaarxiv icon

Linear-MoE: Linear Sequence Modeling Meets Mixture-of-Experts

Add code
Mar 07, 2025
Viaarxiv icon

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

LASP-2: Rethinking Sequence Parallelism for Linear Attention and Its Hybrid

Add code
Feb 11, 2025
Viaarxiv icon

PRMBench: A Fine-grained and Challenging Benchmark for Process-Level Reward Models

Add code
Jan 07, 2025
Figure 1 for PRMBench: A Fine-grained and Challenging Benchmark for Process-Level Reward Models
Figure 2 for PRMBench: A Fine-grained and Challenging Benchmark for Process-Level Reward Models
Figure 3 for PRMBench: A Fine-grained and Challenging Benchmark for Process-Level Reward Models
Figure 4 for PRMBench: A Fine-grained and Challenging Benchmark for Process-Level Reward Models
Viaarxiv icon

LLaMA-MoE v2: Exploring Sparsity of LLaMA from Perspective of Mixture-of-Experts with Post-Training

Add code
Nov 24, 2024
Viaarxiv icon

CLIP-MoE: Towards Building Mixture of Experts for CLIP with Diversified Multiplet Upcycling

Add code
Sep 28, 2024
Viaarxiv icon