Yesheng Liang

2 accepted papers

2026

DFlash: Block Diffusion for Flash Speculative Decoding

ICML 2026poster

Autoregressive large language models (LLMs) deliver strong performance but require inherently sequential decoding, leading to high inference latency and poor GPU utilization. Speculative decoding mitigates this bottleneck by using a fast draft model whose outputs are verified in parallel by the targ…

Cited by 0 Source ↗Scholar ↗

2026

ParoQuant: Pairwise Rotation Quantization for Efficient Reasoning LLM Inference

Yesheng Liang, Haisheng Chen, Song Han, Zhijian Liu

ICLR 2026poster

Post-training quantization (PTQ) compresses the weights and activations of large language models (LLMs) into low-precision representations to reduce memory footprint and accelerate inference. However, the presence of outliers in weights and activations often leads to large quantization errors and se…

Cited by 0 Source ↗code ↗Scholar ↗