Radix Attention

ConceptMentioned in 1 video

A prefix caching technology supported by SGLang, which uses a block size of 1 for potentially higher cache hit rates compared to frameworks using larger block sizes like 32.

Videos Mentioning Radix Attention

DeepSeek V3, SGLang, and the state of Open Model Inference in 2025 (Quantization, MoEs, Pricing)

DeepSeek V3, SGLang, and the state of Open Model Inference in 2025 (Quantization, MoEs, Pricing)

Latent Space

A prefix caching technology supported by SGLang, which uses a block size of 1 for potentially higher cache hit rates compared to frameworks using larger block sizes like 32.