Concept
Round-to-nearest
Round-to-nearest (RTN) is the simplest quantization scheme: map each value to the closest representable level in a fixed grid, breaking ties toward the even level. It is the quantizer equivalent of Python's round(), applied elementwise to a tensor after you pick a scale…
The rest of “Round-to-nearest” is a premium feature: every concept in the library gets a precise, practitioner-focused write-up like this one, cross-linked straight from the paper summaries that use it.
Log in to unlock→