PanGu-$\alpha$: Large-scale Autoregressive Pretrained Chinese Language Models with Auto-parallel Computation

TL;DR

This paper introduces PanGu-$\alpha$, a large-scale autoregressive Chinese language model with 200 billion parameters, trained on extensive data and optimized with advanced parallelism techniques, demonstrating strong few-shot and zero-shot NLP performance.

Contribution

The paper presents the development and training of PanGu-$\alpha$, a 200-billion-parameter Chinese language model utilizing a novel auto-parallel training strategy on a large AI cluster.

Findings

PanGu-$\alpha$ achieves superior few-shot and zero-shot NLP task performance.

Efficient training of 200-billion-parameter model using MindSpore auto-parallelism.

Model generalizes well across diverse Chinese NLP tasks.

Abstract

Large-scale Pretrained Language Models (PLMs) have become the new paradigm for Natural Language Processing (NLP). PLMs with hundreds of billions parameters such as GPT-3 have demonstrated strong performances on natural language understanding and generation with \textit{few-shot in-context} learning. In this work, we present our practice on training large-scale autoregressive language models named PanGu-$\alpha$, with up to 200 billion parameters. PanGu-$\alpha$ is developed under the MindSpore and trained on a cluster of 2048 Ascend 910 AI processors. The training parallelism strategy is implemented based on MindSpore Auto-parallel, which composes five parallelism dimensions to scale the training task to 2048 processors efficiently, including data parallelism, op-level model parallelism, pipeline model parallelism, optimizer model parallelism and rematerialization. To enhance the generalization ability of PanGu-$\alpha$, we collect 1.1TB high-quality Chinese data from a wide range of domains to pretrain the model. We empirically test the generation ability of PanGu-$\alpha$ in various scenarios including text summarization, question answering, dialogue generation, etc. Moreover, we investigate the effect of model scales on the few-shot performances across a broad range of Chinese NLP tasks. The experimental results demonstrate the superior capabilities of PanGu-$\alpha$ in performing various tasks under few-shot or zero-shot settings.