LLM-FP4: 4-Bit Floating-Point Quantized Transformers

Abstract

We propose LLM-FP4 for quantizing both weights and activations in large language models (LLMs) down to 4-bit floating-point values, in a post-training manner. Existing post-training quantization (PTQ) solutions are primarily integer-based and struggle with bit widths below 8 bits. Compared to integer quantization, floating-point (FP) quantization is more flexible and can better handle long-tail or bell-shaped distributions, and it has emerged as a default choice in many hardware platforms. One characteristic of FP quantization is that its performance largely depends on the choice of exponent bits and clipping range. In this regard, we construct a strong FP-PTQ baseline by searching for the optimal quantization parameters. Furthermore, we observe a high inter-channel variance and low intra-channel variance pattern in activation distributions, which adds activation quantization difficulty. We recognize this pattern to be consistent across a spectrum of transformer models designed for diverse tasks, such as LLMs, BERT, and Vision Transformer models. To tackle this, we propose per-channel activation quantization and show that these additional scaling factors can be reparameterized as exponential biases of weights, incurring a negligible cost. Our method, for the first time, can quantize both weights and activations in the LLaMA-13B to only 4-bit and achieves an average score of 63.1 on the common sense zero-shot reasoning tasks, which is only 5.8 lower than the full-precision model, significantly outperforming the previous state-of-the-art by 12.7 points. Code is available at: https://github.com/nbasyl/LLM-FP4.

EMNLP 2023 Main Conference

References in corpus (9)

• LLaMA: Open and Efficient Foundation Language Models
• Scaling Laws for Neural Language Models
• LLM.int8(): 8-bit Matrix Multiplication for Transformers at Scale
• Towards Accurate Post-Training Quantization for Vision Transformer
• FP8 Formats for Deep Learning
• EasyQuant: Post-training Quantization via Scale Optimization
• Towards Efficient Post-training Quantization of Pre-trained Language Models
• LLM-QAT: Data-Free Quantization Aware Training for Large Language Models
• Integer or Floating Point? New Outlooks for Low-Bit Quantization on Large Language Models