Unbacked SymInts PyTorch Developer podcast

Artwork

Tech PyTorch Edward Yang Team PyTorch Deep Learning Machine Learning

内容由PyTorch, Edward Yang, and Team PyTorch提供。所有播客内容（包括剧集、图形和播客描述）均由 PyTorch, Edward Yang, and Team PyTorch 或其播客平台合作伙伴直接上传和提供。如果您认为有人在未经您许可的情况下使用您的受版权保护的作品，您可以按照此处概述的流程进行操作https://zh.player.fm/legal。

PyTorch Developer Podcast « »
Unbacked SymInts

3y ago 21:31

分享

MP3•单集首页

内容由PyTorch, Edward Yang, and Team PyTorch提供。所有播客内容（包括剧集、图形和播客描述）均由 PyTorch, Edward Yang, and Team PyTorch 或其播客平台合作伙伴直接上传和提供。如果您认为有人在未经您许可的情况下使用您的受版权保护的作品，您可以按照此处概述的流程进行操作https://zh.player.fm/legal。

This podcast goes over the basics of unbacked SymInts. You might want to listen to this one before listening to https://pytorch-dev-podcast.simplecast.com/episodes/zero-one-specialization Some questions we answer (h/t from Gregory Chanan):

- Are unbacked symints only for export? Because otherwise I could just break / wait for the actual size. But maybe I can save some retracing / graph breaks perf if I have them too? So the correct statement is "primarily" for export?

- Why am I looking into the broadcasting code at all? Naively, I would expect the export graph to be just a list of ATen ops strung together. Why do I recurse that far down? Why can't I annotate DONT_TRACE_ME_BRO?

- How does 0/1 specialization fit into this? I understand we may want to 0/1 specialize in a dynamic shape regime in "eager" mode (is there a better term?), but that doesn't seem to matter for export?

- So far we've mainly been talking about how to handle our own library code. There is a worry about pushing complicated constraints downstream, similar to torchscript. What constraints does this actually push?

… continue reading

83集单集

#Tech #PyTorch #Edward Yang #Team PyTorch #Deep Learning #Machine Learning

Artwork

Unbacked SymInts

PyTorch Developer Podcast

26 subscribers

published 3y ago

分享

MP3•单集首页

内容由PyTorch, Edward Yang, and Team PyTorch提供。所有播客内容（包括剧集、图形和播客描述）均由 PyTorch, Edward Yang, and Team PyTorch 或其播客平台合作伙伴直接上传和提供。如果您认为有人在未经您许可的情况下使用您的受版权保护的作品，您可以按照此处概述的流程进行操作https://zh.player.fm/legal。

This podcast goes over the basics of unbacked SymInts. You might want to listen to this one before listening to https://pytorch-dev-podcast.simplecast.com/episodes/zero-one-specialization Some questions we answer (h/t from Gregory Chanan):

- Are unbacked symints only for export? Because otherwise I could just break / wait for the actual size. But maybe I can save some retracing / graph breaks perf if I have them too? So the correct statement is "primarily" for export?

- Why am I looking into the broadcasting code at all? Naively, I would expect the export graph to be just a list of ATen ops strung together. Why do I recurse that far down? Why can't I annotate DONT_TRACE_ME_BRO?

- How does 0/1 specialization fit into this? I understand we may want to 0/1 specialize in a dynamic shape regime in "eager" mode (is there a better term?), but that doesn't seem to matter for export?

- So far we've mainly been talking about how to handle our own library code. There is a worry about pushing complicated constraints downstream, similar to torchscript. What constraints does this actually push?

… continue reading

83集单集

#Tech #PyTorch #Edward Yang #Team PyTorch #Deep Learning #Machine Learning

Todos los episodios

×

欢迎使用Player FM

Player FM正在网上搜索高质量的播客，以便您现在享受。它是最好的播客应用程序，适用于安卓、iPhone和网络。注册以跨设备同步订阅。

收听超过500个主题

边探索边听这个节目