探索用于端到端十亿像素深度学习的共享内存架构
Exploring shared memory architectures for end-to-end gigapixel deep learning
AI总结：
本文评估Apple M1 Ultra的128 GB统一内存架构，用轻量改进U-Net直接训练十亿像素WSI组织分割模型，证明端到端全切片深度学习可行。
摘要：
在GPU硬件进步的推动下，深度学习已在医学成像领域取得长足进展。开发新模型的一个主要限制是训练期间GPU内存资源饱和。在计算病理学中尤其如此，因为那里的图像通常包含超过10亿个像素。由于硬件限制，这些病理图像传统上会被划分为小补丁，以便进行深度学习。在本研究中，我们探索Apple, Inc.最近发布的M1 Ultra片上系统（SoC）上的共享GPU/CPU内存架构是否能够提供一种解决方案。这些价格可负担的系统（低于5000美元）可访问128 GB统一内存（配备M1 Ultra SoC的Mac Studio）。作为十亿像素深度学习的概念验证，我们在全切片图像（WSIs）的十亿像素区域上从背景中识别组织。模型是一个改进版U-Net（4492个参数），利用了大卷积核和高步幅。M1 Ultra SoC能够直接在十亿像素图像（16000×64000像素，10.24亿像素）上训练该模型，批量大小为1，该过程使用超过100 GB统一内存，并使用Tensorflow 2/Keras达到每批平均1分21秒的速度。正如预期，模型收敛并取得0.989 ± 0.005的高Dice分数。训练到该阶段经历4940步，耗时111小时24分钟。其他高RAM GPU（例如NVIDIA A100，作为可商业获得的最大型号具有80 GB，约15000美元）尚未广泛可用（在Amazon Web Services上仅面向部分地区提供预览，8个一组的价格为每小时40.96美元）。本研究是利用常见网络架构实现面向WSI整体的端到端深度学习的有希望一步。
英文摘要：
Deep learning has made great strides in medical imaging, enabled by hardware advances in GPUs. One major constraint for the development of new models has been the saturation of GPU memory resources during training. This is especially true in computational pathology, where images regularly contain more than 1 billion pixels. These pathological images are traditionally divided into small patches to enable deep learning due to hardware limitations. In this work, we explore whether the shared GPU/CPU memory architecture on the M1 Ultra systems-on-a-chip (SoCs) recently released by Apple, Inc. may provide a solution. These affordable systems (less than \$5000) provide access to 128 GB of unified memory (Mac Studio with M1 Ultra SoC). As a proof of concept for gigapixel deep learning, we identified tissue from background on gigapixel areas from whole slide images (WSIs). The model was a modified U-Net (4492 parameters) leveraging large kernels and high stride. The M1 Ultra SoC was able to train the model directly on gigapixel images (16000$\times$64000 pixels, 1.024 billion pixels) with a batch size of 1 using over 100 GB of unified memory for the process at an average speed of 1 minute and 21 seconds per batch with Tensorflow 2/Keras. As expected, the model converged with a high Dice score of 0.989 $\pm$ 0.005. Training up until this point took 111 hours and 24 minutes over 4940 steps. Other high RAM GPUs like the NVIDIA A100 (largest commercially accessible at 80 GB, $\sim$\$15000) are not yet widely available (in preview for select regions on Amazon Web Services at \$40.96/hour as a group of 8). This study is a promising step towards WSI-wise end-to-end deep learning with prevalent network architectures.
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。
