Birbla

Join the discussion

Write your take first — we'll ask for email only when you're ready to publish.

Sign in
  • Hacker News
  • Great news, has made low memory bandwidth model usage so much nicer.
    by sarjann
  • vllm PR for DFlash2: https://github.com/vllm-project/vllm/pull/52816
    by verdverm
  • llama.cpp PR https://github.com/ggml-org/llama.cpp/pull/27342
    by cogman10
  • I'm getting around 27 tokens per second decode using vLLM + Qwen 3.8 27b nvfp4 + DFlash 2 on the DGX Spark.
    by adefa
  • I am working this out with Fable right now, for getting this running on my DGX spark homelab; it mentioned that there might be issues with the 'optimized LM-head restrictions' that unsloth NVFP4 ships with. Have you had any issues here?

    Are you trying this with vLLM? Or a different engine? I am getting about 15 tok/s on my spark on my current setup using the 0.26 nvidia vLLM image and MTP.

    by pdntspa
  • Amazing tech

    > An agent writes in an afternoon what a chatbot writes in a month

    But can you just.. not.

    Your tech is so good, it speaks for itself. Don't ruin that.

    by hypfer
  • Watch the video carefully. DFlash2's tool call fails on python syntax.

    Usually models in this class nail things like that 1 shot, which the other side did.

    I don't know the cause. It may be nothing. But I'd like to see the model doing something where its path is a bit more constrained, to help out rule out such oddities.

    by ilc
  • DFlash is lossless so this would be a bug in the implementation if it is indeed a regression against the target model.
    by zackangelo