• 2 Posts
  • 20 Comments
Joined 1 year ago
cake
Cake day: July 22nd, 2023

help-circle
  • I tried Mistral Nemo 12B instruct this morning. It’s actually quite good. I’d say it’s close to dolphin mistral 8x7B which is a monster in size and very smart, about 45 or 50GB. So I’d say Arli is a good deal Mistral Nemo 12B for 4 or $5 per month and privacy so they claim.

    If you don’t mind logging for some questions, you can get access to very good or if not the best models at lmsys.org without monetary cost. Just go to the “Arena”. This is where you contribute with your blind evaluation by voting which of two is better. I often get models like 4o and sonnet 3.5 by Anthropic, google’s best, etc., and at other times many good 70B models. You see two answers at once and vote your favorite between the two. In return, you get “free” access.

    Be careful with AMD GPUs as they are not as well supported for local AI. However, support is gaining ground. Some people are doing it but it takes effort and hassle, from what I’ve read.


  • I know that people are using P40 and P100 GPUs. These are outdated but still work with some software stacks / applications. The P40 GPU, once very cheap for the amount of VRAM, is no longer as cheap as it was probably because folks have been picking them up for inference.

    I’m getting a lot done with an NVidia GTX 1080 which only has 8GB VRAM. I can run a quant of dolphin Mixtral 7x8B and it works well enough. It takes minutes to load, almost too long for me, but after that I get 3-5 TPS with some acceptable delay between questions.

    I can even run Miqu quants at 2 or 3 bits. It’s super smart even at these low quant levels.

    llama 3.1 8B runs great with this 1080 8BG GPU at 4_K_M and also 5 or 6_K_M. I believe I can run gemma 9B f16 at 8 bpw.



  • The community can only read the source code, as of yet. All of the source code has been provided by a set of internal developers.

    The fact that it is open source means that, if somehow two malware elements have made it into the source code, then someone will eventually report it. But this doesn’t mean that two malware elements cannot be there right now.

    These two malware hits on total virus scan should be communicated to the developers.









  • Let’s say I have a favorite sport and there exists a sub_ named: r/.

    Let’s also say there already exits a Lemmy community and that community is struggling to get off the ground: !@lemmy.world

    I can see a value add if your project directly helps !@lemmy.world get started; but I don’t see how it does. If anything wouldn’t your project compete with !@lemmy.world and therefore hinder it?

    It might be different if your project directly tied r/ to !@lemmy.world but it doesn’t.









  • Thanks for the replies. So I guess USENET had/has an advantage here, as all USENET servers replicate “all” newsgroups automatically. To the extent that one server exists, the newsgroup lives on regardless of its origination point. In that sense, the collective work of all contributors is not lost until the retention date passes.

    The ActivityPub proposal mentioned by @chris seems to be a good enough equivalent, at least for communities that are shared.