The Era of 1-bit LLMs: All Large Language Models are in 1.58 Bits
CoRR(2024)
摘要
Recent research, such as BitNet, is paving the way for a new era of 1-bit
Large Language Models (LLMs). In this work, we introduce a 1-bit LLM variant,
namely BitNet b1.58, in which every single parameter (or weight) of the LLM is
ternary -1, 0, 1. It matches the full-precision (i.e., FP16 or BF16)
Transformer LLM with the same model size and training tokens in terms of both
perplexity and end-task performance, while being significantly more
cost-effective in terms of latency, memory, throughput, and energy consumption.
More profoundly, the 1.58-bit LLM defines a new scaling law and recipe for
training new generations of LLMs that are both high-performance and
cost-effective. Furthermore, it enables a new computation paradigm and opens
the door for designing specific hardware optimized for 1-bit LLMs.
更多查看译文
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要