<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>ComfyUI on My learning and diary</title>
    <link>https://jackliusr.github.io/tags/comfyui/</link>
    <description>Recent content in ComfyUI on My learning and diary</description>
    <generator>Hugo -- gohugo.io</generator>
    <language>en-us</language>
    <lastBuildDate>Mon, 18 May 2026 06:00:00 +0800</lastBuildDate><atom:link href="https://jackliusr.github.io/tags/comfyui/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Working Around ROCm PyTorch Replacement Issues with uv and ComfyUI</title>
      <link>https://jackliusr.github.io/posts/2026/05/working-around-rocm-pytorch-replacement-issues-with-uv-and-comfyui/</link>
      <pubDate>Mon, 18 May 2026 06:00:00 +0800</pubDate>
      
      <guid>https://jackliusr.github.io/posts/2026/05/working-around-rocm-pytorch-replacement-issues-with-uv-and-comfyui/</guid>
      <description>Introduction When working with AMD GPUs and ROCm-based AI workloads, one common issue appears when using uv for Python dependency management.
 The problem becomes especially visible when setting up projects like ComfyUI on Linux with ROCm-enabled PyTorch builds.
 Although ROCm-specific wheels are manually installed, running commands such as uv add or dependency synchronization may silently replace ROCm-enabled packages with standard PyPI versions that only support CUDA.
 This leads to broken GPU acceleration and unexpected runtime failures.</description>
    </item>
    
  </channel>
</rss>
