FP8-BERT: Post-Training Quantization for Transformer
This addresses deployment efficiency for NLP applications, but it is incremental as it builds on existing quantization and FP8 formats.
The paper tackles the problem of high memory and inference costs for Transformer models like BERT by proposing FP8-based post-training quantization, showing that it achieves accuracy comparable to full-precision models on GLUE and SQuAD datasets.
Transformer-based models, such as BERT, have been widely applied in a wide range of natural language processing tasks. However, one inevitable side effect is that they require massive memory storage and inference cost when deployed in production. Quantization is one of the popularized ways to alleviate the cost. However, the previous 8-bit quantization strategy based on INT8 data format either suffers from the degradation of accuracy in a Post-Training Quantization (PTQ) fashion or requires an expensive Quantization-Aware Training (QAT) process. Recently, a new numeric format FP8 (i.e. floating-point of 8-bits) has been proposed and supported in commercial AI computing platforms such as H100. In this paper, we empirically validate the effectiveness of FP8 as a way to do Post-Training Quantization without significant loss of accuracy, with a simple calibration and format conversion process. We adopt the FP8 standard proposed by NVIDIA Corp. (2022) in our extensive experiments of BERT variants on GLUE and SQuAD v1.1 datasets, and show that PTQ with FP8 can significantly improve the accuracy upon that with INT8, to the extent of the full-precision model.