5 papers
TaskPress: Query-Agnostic KV Cache Compression via Task-Guided Pruning
Wonpyo Park, Seung-won Hwang
Long-context inference with large language models is constrained by the linear growth of the key-value cache to sequence length. While pruning offers mitigation, prevailing methods…
Activation Quantization of Vision Encoders Needs Prefixing Registers
Seunghyeon Kim, Taesun Yeom, Jinho Kim +3
Large pretrained vision encoders are central to multimodal intelligence, powering applications from on-device vision processing to vision-language models. Since these applications…
GuidedQuant: Large Language Model Quantization via Exploiting End Loss Guidance
Jinuk Kim, Marwa El Halabi, Wonpyo Park +5
Post-training quantization is a key technique for reducing the memory and inference latency of large language models by quantizing weights and activations without requiring retrain…
Rethinking Pruning Large Language Models: Benefits and Pitfalls of Reconstruction Error Minimization
Sungbin Shin, Wonpyo Park, Jaeho Lee +1
This work suggests fundamentally rethinking the current practice of pruning large language models (LLMs). The way it is done is by divide and conquer: split the model into submodel…
Prefixing Attention Sinks can Mitigate Activation Outliers for Large Language Model Quantization
Seungwoo Son, Wonpyo Park, Woohyun Han +2
Despite recent advances in LLM quantization, activation quantization remains to be challenging due to the activation outliers. Conventional remedies, e.g., mixing precisions for di…