LLM.int8(): 8-Bit Matrix Multiplication for Transformers at Scale
Younes Belkada, Tim Dettmers, Mike Lewis, Luke Zettlemoyer
Topics & Concepts
Scale (ratio)TransformerElectronic engineeringComputer scienceMathematicsMatrix (chemical analysis)AlgorithmMultiplication (music)Control theory (sociology)Electrical engineeringEngineeringCryptography and Residue ArithmeticLow-power high-performance VLSI designParallel Computing and Optimization Techniques