itt i document my custom usage of local ai models to make high quality gigas
I go to the Buddha as my refuge. I go to the Dhamma – The Teachings, as my Refuge. I go to the Sangh
>current set up
i got a epic custom model of chatgpt 1 image gen running on comfyui(a node front end that allows for multiple forms of prompting like tag based clip prompting and t5 prompting plus camera tracking prompting and masking)
current result:https://u.pone.rs/tqktzuyz.png
will report more progress
I go to the Buddha as my refuge. I go to the Dhamma – The Teachings, as my Refuge. I go to the Sangh
I go to the Buddha as my refuge. I go to the Dhamma – The Teachings, as my Refuge. I go to the Sangh
while most cacas tend to ai generate their gigas with cloud models on le hecken web(like sora or whatever the latest free ai image generator is) i have decided to run a local model on comfy ui and actually learn how ai models work.
>current set up
i got a epic custom model of chatgpt 1 image gen running on comfyui(a node front end that allows for multiple forms of prompting like tag based clip prompting and t5 prompting plus camera tracking prompting and masking)
current result:https://u.pone.rs/tqktzuyz.png
will report more progress
wew, anyways what were you saying?
I go to the Buddha as my refuge. I go to the Dhamma – The Teachings, as my Refuge. I go to the Sangh
yeah i noticed that, i don't wanna spend money so i'm locally hosting it and forcing myself to learn how ai images are made.nothing web based is free theres always a nigger credits system or limited generations like with veo or hailuo or pixverse
currently reading le science paper on ai image schedulers(basically how the ai paints noise on to make a image and then refines said noise) Karras(which isn't really used much it seems) is interesting cuz it calculates a curve coeffiecent for how much noise should be painted on via painting large noise data and then smaller details later on
I go to the Buddha as my refuge. I go to the Dhamma – The Teachings, as my Refuge. I go to the Sangh
I go to the Buddha as my refuge. I go to the Dhamma – The Teachings, as my Refuge. I go to the Sangh
wew, anyways what were you saying?
https://soyjakwiki.org/AI#Running_an_AI_locallybtw has anyone else tried this or only me?
probably if they read the ki
wew, anyways what were you saying?
same but im probably too stupid to do thisyeah i noticed that, i don't wanna spend money so i'm locally hosting it and forcing myself to learn how ai images are made.Show quoted text
nothing web based is free theres always a nigger credits system or limited generations like with veo or hailuo or pixverse
currently reading le science paper on ai image schedulers(basically how the ai paints noise on to make a image and then refines said noise) Karras(which isn't really used much it seems) is interesting cuz it calculates a curve coeffiecent for how much noise should be painted on via painting large noise data and then smaller details later on
I go to the Buddha as my refuge. I go to the Dhamma – The Teachings, as my Refuge. I go to the Sangh
implying modern soyim read the ki and haven't found the boorum from tikki talkhttps://soyjakwiki.org/AI#Running_an_AI_locallyShow quoted text
btw has anyone else tried this or only me?
probably if they read the ki
I go to the Buddha as my refuge. I go to the Dhamma – The Teachings, as my Refuge. I go to the Sangh
tbh most of it is just high school math, ai image generation is just a fine tuned markhov chain that paints random noise till it makes something that it was tuned to dosame but im probably too stupid to do thisShow quoted text
yeah i noticed that, i don't wanna spend money so i'm locally hosting it and forcing myself to learn how ai images are made.
currently reading le science paper on ai image schedulers(basically how the ai paints noise on to make a image and then refines said noise) Karras(which isn't really used much it seems) is interesting cuz it calculates a curve coeffiecent for how much noise should be painted on via painting large noise data and then smaller details later on
I go to the Buddha as my refuge. I go to the Dhamma – The Teachings, as my Refuge. I go to the Sangh
I go to the Buddha as my refuge. I go to the Dhamma – The Teachings, as my Refuge. I go to the Sangh
i shall call this piece gigas raping
I go to the Buddha as my refuge. I go to the Dhamma – The Teachings, as my Refuge. I go to the Sangh
I go to the Buddha as my refuge. I go to the Dhamma – The Teachings, as my Refuge. I go to the Sangh
must save.. booru..
I go to the Buddha as my refuge. I go to the Dhamma – The Teachings, as my Refuge. I go to the Sangh
this is scary and incomphresible, ussed the dpms++ sampler
https://medium.com/data-science/understanding-diffusion-probabilistic-models-dpms-1940329d6048
the actual math is lost but from intution it like takes the pieces it sees and does a kinda reverse function operation? it's like if you were to redo the steps of painting on a image i think
I go to the Buddha as my refuge. I go to the Dhamma – The Teachings, as my Refuge. I go to the Sangh
what nigger boy want some sheep MILKWhy do xou have that as xoure pfp
I go to the Buddha as my refuge. I go to the Dhamma – The Teachings, as my Refuge. I go to the Sangh
current result: need to be more detailed probably might be forced to change models doe