PV-Tuning

Yandex researchers develop new methods for compressing large language models, cutting AI deployment costs by up to 8 times