Bringing Up DeepSeek-V4-Flash on AMD MI300X(fergusfinn.com)
84 pointsby kkm9 hours ago6 comments
  • maCDzP6 hours ago
    I train on AMD MI250X and managed to get Gemma 4 31B to work - but it took a lot of work on the software side.
    • kkm6 hours ago
      This is very interesting, planning to write about it?
  • kkm8 hours ago
    Also the vllm patch accompanying the blogpost: https://github.com/doublewordai/vllm-amd-blog-doubleword
  • mezark8 hours ago
    We at doubleword are bullish for AMD for low-interactivity inference - it does just take a bigger lift on the software side...
    • brcmthrowaway6 hours ago
      Are you long AMD?
      • latchkey7 minutes ago
        Interesting that you ask that as AMD hits another ATH.
  • benlm8 hours ago
    Nice work! Would DeepSeek V4 Pro on 8xMI300X work with these patches?
  • latchkey3 hours ago
    Nice work and thanks for being a customer.

    (CEO Hot Aisle)

  • jing099282 hours ago
    [flagged]