USearch introduces FP8 (8-bit floating point) support for vector search and KV-caching, enabling more efficient memory usage and faster computations. The implementation allows for reduced storage requirements while maintaining search accuracy through quantization techniques.
FP4 is a 4-bit floating point format that represents a significant reduction from traditional 32-bit and 64-bit floating point standards. This compact format enables more efficient storage and computation in resource-constrained environments like edge devices and AI accelerators.
This post examines the NF4 4-bit floating point format and higher precision analogs used for quantizing LLM weights. NF4 and FP4 are common 4-bit data types in bitsandbytes, often found in weights downloaded from Hugging Face.