xcjs

@xcjs@programming.dev

• ⁨Comment⁩ on ⁨[JS Required] MiniMax M1 model claims Chinese LLM crown from DeepSeek - plus it's true open-source⁩ ⁨⁨1⁩ ⁨week⁩ ago⁩:

  That’s not how distillation works if I understand what you’re trying to explain.

  If you distill model A to a smaller model, you just get a smaller version of model A with the same approximate distribution curve of parameters, but fewer of them. You can’t distill Llama into Deepseek R1.

  I’ve been able to run distillations of Deepseek R1 up to 70B, and they’re all censored still. There is a version of Deepseek R1 “patched” with western values called R1-1776 that will answer topics censored by the Chinese government, however.

• ⁨Comment⁩ on ⁨That's all folks, Plex is starting to charge for sharing⁩ ⁨⁨1⁩ ⁨month⁩ ago⁩:

  The client is open source and can be administered using the open source Headscale server. I use it with Keycloak as an auth gateway.

• ⁨Comment⁩ on ⁨Steam doesn’t want to pay arbitration fees, tells gamers to sue instead⁩ ⁨⁨8⁩ ⁨months⁩ ago⁩:

  If it wasn’t better than that, no company would want arbitration cases.