so I have a weird idea, I want to make (train?) an ai model thing so I have a prompt (directions for a song) and you get back a text response with a song that heavily resembles say yeat’s songs
so basically I kind of want to have a mix of openai’s models (or well like how gpt 3 functions) with a songwriter and have it read multiple files with a songwriter’s songs and the model learns (is trained on?) the patterns of the song and now can create yeat songs based on a user’s prompt
so how would I do this? not very familiar with ai so idk how to do this or if text based models even work like how I think they work
it’s only gonna be used personally, not distributed, and this question also extends to basically creating your own ai model based on information you give
@doxr do you have $$$? If so, then get ChatGPT Plus and make a chatbot. If not, try to find a hugging face model that does what/kind of what you need to do. Also, can’t you just give a different prompt? Also, 2, if you trained a bot, you would need a spare GPU and a lot of money to pay for the training cost on the various websites. (Ff I remember correctly, X, formerly Twitter, charges around $60k)