pytorch/serve

Accelerated preprocessing

Open

#1,546 创建于 2022年4月1日

在 GitHub 查看
 (1 评论) (3 反应) (0 负责人)Java (3,844 star) (790 fork)batch import
enhancementhelp wanted

描述

This recent PR by @min-jean-cho https://github.com/pytorch/serve/pull/1545 has made me realize how much performance we're leaving on the table when it comes to acceleration of preprocessing operations. We could more easily support more torchvision backends, faster tokenizers and add more hardware specific configurations to make it for developers to pick and choose.

This will also help our #1457 efforts since there we'd only considered so far performance of torchserve and inference()

  • Accelerated decoding
  • Accelerated preprocessing ops
  • Graph mode accelerated processing ops - like DALI

贡献者指南

技术栈
python
领域
performancemachine learning
议题类型
feature
难度面向新贡献者的预计实现难度，1 表示很小改动，5 表示专家级工作。
3
预计时间有经验贡献者完成调查、实现、测试并准备 pull request 的粗略时间范围。
1-2 days
活动状态议题当前的可参与程度：新鲜、活跃、陈旧、阻塞或等待维护者输入。
stale
清晰度议题是否清楚说明期望改动、验收标准和下一步。
mostly clear
前置要求
Basic knowledge of PyTorch ServePython programmingFamiliarity with torchvision
新手友好度1-100 的估计分数，表示该议题对首次贡献者的友好程度。
40
研究方向
Review the linked PR #1545 for acceleration ideas. Investigate the current preprocessing pipeline in PyTorch Serve, focusing on image and text preprocessing. Consider integrating torchvision's acceleration features and hardware specific optimizations. Examine existing code in the `serve/` directory for potential entry points to implement accelerated decoding and preprocessing ops.