Changsheng Zhao

MobileLLM-R1: Exploring the Limits of Sub-Billion Language Model Reasoners with Open Training Recipes

DepthLM: Metric Depth from Vision Language Models

STEM: SCALING TRANSFORMERS WITH EMBEDDING MODULES

WiniQ: Accelerating Quantization-Aware Training of LLMs around Saddle Points

Know When to Fold 'Em: Predicting an LLM-Judge for Efficient but Performant Inference

ParetoQ: Improving Scaling Laws in Extremely Low-bit LLM Quantization

Agent-as-a-Judge: Evaluate Agents with Agents

LongVU: Spatiotemporal Adaptive Compression for Long Video-Language Understanding

SpinQuant: LLM Quantization with Learned Rotations

LongVU: Spatiotemporal Adaptive Compression for Long Video-Language Understanding

Agent-as-a-Judge: Evaluating Agents with Agents

Basel: Target-Aware Basis Selection for Language Models

Advancing Energy Efficiency in On-Device Streaming Speech Recognition