• 0 Posts
  • 18 Comments
Joined 1 year ago
cake
Cake day: June 9th, 2023

help-circle








  • Well unlike your ass, I appreciate the nuance of a good performance. But i know what you mean.

    GPU time, while cheaper than a voice actor, is still a bit spendy though. And you then you also have the various copyright/licensing “issues” associated with AI content, companies may be a bit hesitant to go all in on producing books like that. Makes more sense for someone like Amazon/Audible and less sense for someone like spotify.

    Besides, most audio books exist already so that really only applies to newer titles.




  • I’m not defending them, just saying that it’s foolish for an enduser to expect anything different when they already don’t pay musicians and that is the primary content on their platform.

    I can pretty much guarantee the average user would complain way more about the quality of simple TTS than they would the time limit. It would likely be a much bigger PR issue for them. AI generated TTS would probably be good enough for most but that is just another cost.

    Regardless, the licensing involved with book publishers wouldn’t allow them to just produce their own audiobooks like that. So it is not really as simple as “just a choice”.







  • If you just want to use a local llm, using something like gpt4all is probably the easiest. Oobabooga or llama.cpp for a more advanced route.

    I use ollama with llama3 on my macbook with open-webui and it works real nice. Mistral7b is another one I like. On my PC I have been using oobabooga with models I get from huggingface and I use it as an api for hobby projects.

    I have never trained models, I don’t have the vram. My GPU is pretty old so I just use these for random gamedev and webdev projects and for messing around with RP in sillytavern.