5 papers
ThinKV: Thought-Adaptive KV Cache Compression for Efficient Reasoning Models
Akshat Ramachandran, Marina Neseem, Charbel Sakr +3
The long-output context generation of large reasoning models enables extended chain of thought (CoT) but also drives rapid growth of the key-value (KV) cache, quickly overwhelming…
Pretraining Large Language Models with NVFP4
NVIDIA, Felix Abecassis, Anjulie Agrusa +87
Large Language Models (LLMs) today are powerful problem solvers across many domains, and they continue to get stronger as they scale in model size, training set size, and training…
LO-BCQ: Block Clustered Quantization for 4-bit (W4A4) LLM Inference
Reena Elangovan, Charbel Sakr, Anand Raghunathan +1
Post-training quantization (PTQ) is a promising approach to reducing the storage and computational requirements of large language models (LLMs) without additional training cost. Re…
FGMP: Fine-Grained Mixed-Precision Weight and Activation Quantization for Hardware-Accelerated LLM Inference
Coleman Hooper, Charbel Sakr, Ben Keller +4
Quantization is a powerful tool to improve large language model (LLM) inference efficiency by utilizing more energy-efficient low-precision datapaths and reducing memory footprint.…
ESPACE: Dimensionality Reduction of Activations for Model Compression
Charbel Sakr, Brucek Khailany
We propose ESPACE, an LLM compression technique based on dimensionality reduction of activations. Unlike prior works on weight-centric tensor decomposition, ESPACE projects activat…