KV, a low-rank KV cache compression method achieving up to 20x reduction, with the paper selected as a Spotlight at ICML 2026 ...

Introduces a low-rank-based approach to KV cache compression, one of the key bottlenecks in long-context AI; Speeds up ...

Sophisticated AI models tend to require a lot of memory and take up a lot of storage space. One of the ways to reduce that ...

Zcash is building a new consensus layer that keeps mining alive while adding a stake-based finality check. The proposed ...

You don't always need an RTX 5090 to run useful models ...

You can now download Gemma 4 models with quantization-aware training to reduce the amount of mobile memory required to 1GB.

Large language models (LLMs) aren’t actually giant computer brains. Instead, they are massive vector spaces in which the probabilities of tokens occurring in a specific order is encoded. Billions of ...

Abstract: Quantization is a critical technique employed across various research fields for compressing deep neural networks (DNNs) to facilitate deployment within resource-limited environments. This ...

As Large Language Models (LLMs) expand their context windows to process massive documents and intricate conversations, they encounter a brutal hardware reality known as the "Key-Value (KV) cache ...

Atom is an accurate low-bit weight-activation quantization algorithm that combines (1) mixed-precision, (2) fine-grained group quantization, (3) dynamic activation quantization, (4) KV-cache ...