论文标题
FP8格式用于深度学习
FP8 Formats for Deep Learning
论文作者
论文摘要
FP8是加速深度学习训练推论以外的16位格式的自然发展。在本文中，我们提出了一个8位浮点（FP8）二进制互换格式，该格式由两个编码组成-E4M3（4位指数和3位Mantissa）和E5M2（5位指数和2位指数和2位Mantissa）。尽管E5M2遵循IEEE 754惯例代表特殊值的惯例，但E4M3的动态范围是通过不代表无限态，只有一个Mantissa bit-Pattern来扩展的。我们证明了FP8格式对各种图像和语言任务的功效，从而有效地匹配了16位培训课程所达到的质量。我们的研究涵盖了主要的现代神经网络体系结构 - CNN，RNN和基于变压器的模型，使所有超参数从16位基线训练课程中保持不变。我们的培训实验包括大型，最多175b参数，语言模型。我们还检查了使用16位格式训练的语言模型的FP8训练后定量化，以抵抗固定点INT8量化。
FP8 is a natural progression for accelerating deep learning training inference beyond the 16-bit formats common in modern processors. In this paper we propose an 8-bit floating point (FP8) binary interchange format consisting of two encodings - E4M3 (4-bit exponent and 3-bit mantissa) and E5M2 (5-bit exponent and 2-bit mantissa). While E5M2 follows IEEE 754 conventions for representatio of special values, E4M3's dynamic range is extended by not representing infinities and having only one mantissa bit-pattern for NaNs. We demonstrate the efficacy of the FP8 format on a variety of image and language tasks, effectively matching the result quality achieved by 16-bit training sessions. Our study covers the main modern neural network architectures - CNNs, RNNs, and Transformer-based models, leaving all the hyperparameters unchanged from the 16-bit baseline training sessions. Our training experiments include large, up to 175B parameter, language models. We also examine FP8 post-training-quantization of language models trained using 16-bit formats that resisted fixed point int8 quantization.
