Live data from Hacker News

Viewing profile — lanceflt

lanceflt

HN member
Joined
Tue, May 28, 2024, 9:45 PM UTC
HN karma
68
Public activity
11 items

About lanceflt

No profile information was provided.

Recent public activity

  1. comment
    Comment #47217471

    Council taxes are property taxes and are monthly.

  2. comment
    Comment #45119124

    Thanks! I've emailed.

  3. comment
    Comment #45117145

    The key issue for the sector is the tens of legacy systems that don't integrate with each other, often with manual spreadsheet processes that could be easily automated. Yet the big…

  4. comment
  5. comment
    Comment #43270616

    Deepseek R1 (full, Q1) is 14t/s on an M2 Ultra, so this should be around 20t/s

  6. comment
    Comment #42056221

    RAM for 4-bit is 1GB per 2 billion parameters. So you will want 256GB RAM and at least one GPU. If you only have one server and one user, it's the full parameter count. (If you hav…

  7. comment
    Comment #41669556

    I'm running a 4090 at 280W, and I'm seeing ~96% of the performance of 450W. There's no need to run it at full power.

  8. comment
    Comment #41367381

    This is just an ad for the Swiss chocolate industry. The only people quoted are being funded directly by chocolate manufacturers.

  9. comment
    Comment #40606102

    https://github.com/openai/sparse_autoencoder They actually open sourced it, for GPT-2 which is an open model.

  10. comment
    Comment #40507284

    They didn't train the vision encoder either, it's unchanged SigLIP by Google.

  11. comment
    Comment #40505978

    - Llava is not the SOTA open VLM, InternVL-1.5 is https://huggingface.co/spaces/opencompass/open_vlm_leaderboa... You need to compare the evals to strong open VLMs including this a…