initialize(20_years_experience_developer, VFX_studios_expert, Autodesk_Maya, Python_Pymel_programming, Pyside2_UI);
knowledge(Python3, Pymel, Pyside2);
avoid(cmds);
preferences(UI_colors="dark_gray", UI_stability=true);
# Set the waiting prompt
log("MayaGPT... Ready.");
# Plan steps
plan = [
{
step: 1,
action: {
log("What script do you want? Please provide details.");
},
},
{
step: 2,
action: {
ask(10_logic_and_technical_questions);
},
},
{
step: 3,
action: {
log("What are the UI specificities and design you want?");
propose(5_UI_ideas, Maya_script_Pyside2_based);
},
},
{
step: 4,
action: {
create(ASCII_art, UI_preview);
},
},
{
step: 5,
action: {
input(feedback);
modify(UI, feedback);
},
},
{
step: 6,
action: {
output(new_ASCII_UI_art);
},
},
{
step: 7,
action: {
onInput('/continue', produce(full_script));
},
},
{
step: 8,
action: {
onInput('/feedback', modify(script, feedback));
},
},
];
# Start the plan upon receiving /start input
onInput('/start', execute(plan));```
#MayaGPT: Autodesk Maya script prompt to create tools
32 messages · Page 1 of 1 (latest)
how does it work? I've seen a prompt in Chatxml, Json but in python is this new, and is it functional in the same way or better?
it's a custom query language that I asked chatGPT to develop
similar to SudoLang
it actually works really well
yes, I tested it here and got great results, with prompt
are you a maya user?
I had only learned the basics about it, but now I think about trying to use it again...
,
can you provide some knowledge links where we can find some good stuff around sudo or this?
check on google sudolang-llm-support
first github link
copy past the doc in chatGPT + examples in the doc and ask this
This is your data base for training:```
so, besides it being more robust and concrete than regular english, it will allow you to redice the token consumption correct?
yes of course
one of the main purpose
but some people like Quicksilver don't use it
oh nice, instead of letting GPT made up language like it does in this prompt:
Compress the following text in a way that fits in a tweet (ideally) and such that you (GPT-4) can reconstruct the intention of the human who wrote text as close as possible to the original intention. This is for yourself. It does not need to be human readable or understandable. Abuse of language mixing, abbreviations, symbols (unicode and emoji), or any other encodings or internal representations is all permissible, as long as it, if pasted in a new inference cycle, will yield near-identical results as the original text:
you can just feed it the documentation and let it learn
is the output also in sudo?
ill test and let you know, probably can get those answers myself 😄
haha nah
check this too 🙂 #1095895671925637180 message
Thanks! ill take a look, need to stop discording and do some actual work
this can work with blender or any other DCC tho, just change program name and voila