<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/">
  <channel>
    <title>Pytorch on Tech News Feed</title>
    <link>https://news.dhphong.com/tags/pytorch/</link>
    <description>Recent content in Pytorch on Tech News Feed</description>
    <generator>Hugo -- 0.131.0</generator>
    <language>vi</language>
    <lastBuildDate>Thu, 30 Apr 2026 00:02:53 +0700</lastBuildDate>
    <atom:link href="https://news.dhphong.com/tags/pytorch/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>[Google Engineering Blog] Speeding Up AI: Bringing Google Colossus to PyTorch via GCSFS and Rapid Bucket</title>
      <link>https://news.dhphong.com/posts/2026-04-29-google-colossus-pytorch-gcsfs-rapid-bucket/</link>
      <pubDate>Thu, 30 Apr 2026 00:02:53 +0700</pubDate>
      <guid>https://news.dhphong.com/posts/2026-04-29-google-colossus-pytorch-gcsfs-rapid-bucket/</guid>
      <description>Nguồn: Google Engineering Blog
Tóm tắt Google công bố tích hợp Rapid Storage — dựa trên kiến trúc lưu trữ Colossus (cùng hệ thống file system vận hành YouTube và Google Search) — trực tiếp vào PyTorch ecosystem thông qua giao diện chuẩn fsspec. Rapid Bucket là giải pháp high-performance object storage trong dedicated zonal buckets, thay thế legacy REST API bằng persistent gRPC bidirectional streams, đạt throughput 15+ TiB/s, latency dưới 1ms cho random reads/appends, và 20M+ QPS.</description>
    </item>
  </channel>
</rss>
