10 papers
Recovering Low-Bit Failures in Vision-Language-Action Policies via Quantization-Conditioned Reinforcement Learning
Seongmin Park, Jaewon Lee, Yujun Heo +4
Quantization lowers deployment costs of vision-language-action (VLA) policies but can compromise task completion. Across five VLA models and 4,050 real-world trials, we systematica…
BeaconKV: Key-Value Cache Compression Guided by Beacon Queries for Efficient Large Reasoning Model Inference
Janghyeon Kim, Minsoo Kim, Kyuhong Shim +1
Large Reasoning Models (LRMs) achieve superior problem-solving through extended Chain-of-Thought (CoT) generation, but the resulting key-value (KV) cache grows linearly with sequen…
Tangram: Unlocking Non-Uniform KV Cache Compression for Efficient Multi-turn LLM Serving
Hyungmin Kim, Minsoo Kim, Hongseok Kim +1
Multi-turn LLM serving accumulates dialogue history whose Key-Value (KV) cache grows with every turn and every user, quickly exceeding the model weights themselves and making memor…
InfiniPot-V: Memory-Constrained KV Cache Compression for Streaming Video Understanding
Minsoo Kim, Kyuhong Shim, Jungwook Choi +1
Modern multimodal large language models (MLLMs) can reason over hour-long video, yet their key-value (KV) cache grows linearly with time-quickly exceeding the fixed memory of phone…
Saliency-Aware Quantized Imitation Learning for Efficient Robotic Control
Seongmin Park, Hyungmin Kim, Sangwoo Kim +5
Deep neural network (DNN)-based policy models, such as vision-language-action (VLA) models, excel at automating complex decision-making from multi-modal inputs. However, scaling th…
Quantization-Aware Imitation-Learning for Resource-Efficient Robotic Control
Seongmin Park, Hyungmin Kim, Wonseok Jeon +4
Deep neural network (DNN)-based policy models like vision-language-action (VLA) models are transformative in automating complex decision-making across applications by interpreting…
RILQ: Rank-Insensitive LoRA-based Quantization Error Compensation for Boosting 2-bit Large Language Model Accuracy
Geonho Lee, Janghwan Lee, Sukjin Hong +4
Low-rank adaptation (LoRA) has become the dominant method for parameter-efficient LLM fine-tuning, with LoRA-based quantization error compensation (LQEC) emerging as a powerful too…
InfiniPot: Infinite Context Processing on Memory-Constrained LLMs
Minsoo Kim, Kyuhong Shim, Jungwook Choi +1
Handling long input contexts remains a significant challenge for Large Language Models (LLMs), particularly in resource-constrained environments such as mobile devices. Our work ai…
Improving Conversational Abilities of Quantized Large Language Models via Direct Preference Alignment
Janghwan Lee, Seongmin Park, Sukjin Hong +3
The rapid advancement of large language models (LLMs) has facilitated their transformation into conversational chatbots that can grasp contextual nuances and generate pertinent sen…
Enhancing Computation Efficiency in Large Language Models through Weight and Activation Quantization
Janghwan Lee, Minsoo Kim, Seungcheol Baek +3
Large Language Models (LLMs) are proficient in natural language processing tasks, but their deployment is often restricted by extensive parameter sizes and computational demands. T…