arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2010.14701cs.LGcs.CLcs.CV
收藏

自回归生成模型的缩放定律

Scaling Laws for Autoregressive Generative Modeling

Johns Hopkins University(约翰斯·霍普金斯大学);OpenAI
Tom Henighan, Jared Kaplan, Mor Katz, Mark Chen, Christopher Hesse, Jacob Jackson, Heewoo Jun, Tom B. Brown, Prafulla Dhariwal, Scott Gray, Chris Hallacy, Benja… 展开作者

Tom Henighan, Jared Kaplan, Mor Katz, Mark Chen, Christopher Hesse, Jacob Jackson, Heewoo Jun, Tom B. Brown, Prafulla Dhariwal, Scott Gray, Chris Hallacy, Benjamin Mann, Alec Radford, Aditya Ramesh, Nick Ryder, Daniel M. Ziegler, John Schulman, Dario Amodei, Sam McCandlish

摘要:

我们确定经验缩放定律的交叉熵损失在四个领域:生成图像建模,视频建模,多模态图像$\leftrightarrow$文本模型,和数学问题的解决。在所有情况下,自回归Transformers的性能随着模型大小和计算预算的增加而平滑地提高,遵循幂律加上常数缩放律。最佳模型大小还取决于幂律的计算预算,其指数在所有数据域中几乎是通用的。交叉熵损失有一个信息论解释为$S($ True $) + D_{\mathrm{KL}}($ True $||$ Model $)$,经验缩放定律建议预测真实数据分布的熵和真实分布与模型分布之间的KL散度。通过这种解释,十亿参数Transformers是下采样到$8\times 8$分辨率的YFCC 100 M图像分布的近乎完美的模型,并且我们可以预测在nats/图像中实现任何给定的可减少损失(即$D_{\mathrm{KL}}$)所需的模型大小。我们发现了一些额外的缩放定律在特定的领域:(a)我们确定了多模态模型中的标题和图像之间的互信息的缩放关系,并展示了如何回答这个问题“是一张图片值得一千字?";(b)在数学问题解决的情况下,当外推超出训练分布时,我们确定了模型性能的缩放定律;(c)我们微调了ImageNet分类的生成图像模型,并发现分类损失和错误率的平滑缩放,即使生成损失趋于平稳。总之,这些结果加强了缩放定律对神经网络性能(包括下游任务)具有重要影响的情况。

英文摘要:

We identify empirical scaling laws for the cross-entropy loss in four domains: generative image modeling, video modeling, multimodal image$\leftrightarrow$text models, and mathematical problem solving. In all cases autoregressive Transformers smoothly improve in performance as model size and compute budgets increase, following a power-law plus constant scaling law. The optimal model size also depends on the compute budget through a power-law, with exponents that are nearly universal across all data domains. The cross-entropy loss has an information theoretic interpretation as $S($True$) + D_{\mathrm{KL}}($True$||$Model$)$, and the empirical scaling laws suggest a prediction for both the true data distribution's entropy and the KL divergence between the true and model distributions. With this interpretation, billion-parameter Transformers are nearly perfect models of the YFCC100M image distribution downsampled to an $8\times 8$ resolution, and we can forecast the model size needed to achieve any given reducible loss (ie $D_{\mathrm{KL}}$) in nats/image for other resolutions. We find a number of additional scaling laws in specific domains: (a) we identify a scaling relation for the mutual information between captions and images in multimodal models, and show how to answer the question "Is a picture worth a thousand words?"; (b) in the case of mathematical problem solving, we identify scaling laws for model performance when extrapolating beyond the training distribution; (c) we finetune generative image models for ImageNet classification and find smooth scaling of the classification loss and error rate, even as the generative loss levels off. Taken together, these results strengthen the case that scaling laws have important implications for neural network performance, including on downstream tasks.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑