New
kindlingai/exlr8 — Low-bit quantization format for DGX Spark
kindlingai/exlr8 is a quantization format for DGX Spark focused on reducing the amount of expert weights read per token.
kindlingai/exlr8 is a quantization format for DGX Spark focused on reducing the amount of expert weights read per token.
View on GitHub ↗kindlingai/exlr8 is a quantization format for DGX Spark focused on reducing the amount of expert weights read per token.
No matching stories.