Zhefeng Wang

Direction-Magnitude Decoupling for Fast Video Generation with Flow Matching Models

CaliDrop: KV Cache Compression with Query-based Calibration

RA-SpaRC: Robust Adaptation with Sparse Plus Low-Rank Compressors

Adaptive Dual-Granularity Pruning Method for Large Language Models

Accelerating Large Language Model Inference via Speculative Decoding with Progressive Tree Drafting

Semantic-aware Pruning of Large Language Models via Neuron Importance Explanation

Adapprox: Memory Efficient Optimization via Adaptive Randomized Low-Rank Approximation

Efficiently Serving Large Multimodal Models Using EPD Disaggregation

Beware of Calibration Data for Pruning Large Language Models

FISTAPruner: Layer-wise Post-training Pruning for Large Language Models

SinkQ: Accurate 2-bit KV Cache Quantization with Dynamic Sink Tracking

FASP: Fast and Accurate Structured Pruning of Large Language Models

CASD: Enhancing Generation Accuracy via Context-Aware Speculative Decoding