multi-precision

Tag

Cards List
#multi-precision

Recurrent Residual Quantization: A Progressive Multi-Precision Representation for LLMs

arXiv cs.LG · 4d ago Cached

Introduces Recurrent Residual Quantization (RRQ), a post-training quantization framework that enables multiple effective precisions from a single LLM checkpoint via additive residual refinement, improving flexibility and construction speed.

0 favorites 0 likes
← Back to home

Submit Feedback