𞋴𝛂𝛋𝛆

  • 56 Posts
  • 565 Comments
Joined 2 years ago
cake
Cake day: June 9th, 2023

help-circle






  • Gptel is better than ellama. Ollama only runs small junk. Llama.cpp is way better because it will split CPU and GPU and run bigger quantized models, especially large MoEs on a 16 GB GPU.

    I started trying to mess with image-dired today for a few minutes to try and get my training images and caption files to sync and scroll, but didn’t get very far before just tiling… such a noob





  • Essentially he was. This is him saying I have no moral threshold and am available as an influencer to say and do whatever I am paid for. AKA I am still relevant to your agenda… Call me. It is the same as all the big tech CEOs bending to 47 after bending to diversity initiatives. People with no ethics or morals fly the banner of whomever has control and do so in ways the average person is far too stupid to process. 47 is hard pressed for black and minority supporters. For a leader of dogmatic tribal fools, aka an influencer, the vacuum of minority supporters for 47 is an open bounty for a very large paycheck in exchange for ethics especially in the Southeastern USA where any sway over potential Democratic voters is critical for the fascist republican agenda. Nothing is simple or straight forward when a person is at this level of political engagement and wealth.





  • What is creepy about it? The entire story is really about Alice discovering rules and how she likes them so much. That is very much the world of an 8 year old. Not to mention all of the fanciful characters and tropes that were dreamed up.

    There are many academic critiques and studies of the work. It is also central to Open AI’s QKV layers model alignment training. The story is the primary catalyst for creative randomness in terms of internal model thinking in both LLMs and embedding models used in image diffusion. For instance, all of the mechanisms that Alice used to become bigger and smaller are present in diffusion AI models with the caveat that no real person place or thing is present in the prompt. One does not need to bring up the detailed context of the story if one prompts the element with good specificity. A far easier method to play with is to prompt the queen of hearts or Alice as a character in an image using just a foundational base model. It should be quite clear how these images are a bit different in many ways. Those differences are not random and they are persistent across all models. Literary nonsense is the actual randomness that shows up in background objects and clothing in images. Prompting against the abstraction of genre is far more effective than attempting to describe your own details in the prompt with specificity.

    In literature in general, Carroll’s work is the holotype for a genre. Analysing the work speaks to the human experience on many levels. The work has long been appreciated by all ages. So I am a bit baffled about what you find creepy about age in any context such as this. Like what kind of assumptions do you possibly feel grounded in here?