Pixar USD & XSI .dotXSI and XSI Python
These aren't quite fully formed but they are getting close.
Prompt: Generate Pixar USD of a cube at the origin of size 10,10,10
The file doesn't import as a USD import in Blender but it has potential.
Code: Select all
// PixarUSD Cube at the Origin of Size 10,10,10
#usda
def Cube "cube"
{
float3 xformOp[1] = {
"xformOp:translate", 0.0, 0.0, 0.0
};
float3 xformOpScale[1] = {
"xformOp:scale", 10.0, 10.0, 10.0
};
float3x3 xformOpOrient[1] = {
"xformOp:orient", 1.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 1.0
};
float3 xformOpOrder[1] = {
"xformOp:order", 0.0, 1.0, 2.0
};
string type = "cube";
}
Haven't checked that it loads but again a lot of potential.
Code: Select all
// Softimage XSI Cube at the Origin of Size 10,10,10
<?xml version="1.0" encoding="UTF-8"?>
<XSIFile>
<Model Name="cube">
<Properties>
<Transformation>
<Translation>0.0 0.0 0.0 </Translation>
<Rotation>0.0 0.0 0.0 </Rotation>
<Scaling>10.0 10.0 10.0 </Scaling>
</Transformation>
</Properties>
</Model>
</XSIFile>
Updated this prompt so it's no longer a convert from what it had preciously generated. For a much better result.
Prompt: create the a cude in softimage xsi at the origin size 10,101,10 using Python
Code: Select all
# Softimage XSI Python Cube at the Origin of Size 10,10,10
from xsi import xsi
cube = xsi.ActiveSceneRoot.AddPrim('Cube', 'MeshSurface')
cube.Kinematics.Global.Transform.Translation = [0.0, 0.0, 0.0]
cube.Kinematics.Global.Transform.Rotation = [0.0, 0.0, 0.0]
cube.Kinematics.Global.Transform.Scaling = [10.0, 10.0, 10.0]
I wonder if there is a way to convert from binary to ASCII files and the reverse? Ive tried but failed.
Not me: "I Made a Game with ChatGPT... ( In 3 hours using only GPT )"
https://www.youtube.com/watch?v=L6vW0um5XTg
AI Artificial Intelligence OpenAI Chat GPT
Its free for now but I wonder what they will eventually charge?
Sign up for free ( at your own risk ). It asks for your cell phone number, name and email. Don’t enter any private information.
https://chat.openai.com/
https://beta.openai.com/playground
if you ever want to delete your data
https://help.openai.com/en/articles/637 ... my-account
Sharing & Publication PolicyChatGPT continues to suffer from the fundamental problems of current language models:
• They invent facts,
• but present them with great confidence in polished words,
• they offer no assessment of the reliability of the information
• and no source transparency.
• Moreover, large language models can reinforce prejudices and create new ones
ChatGPT is “incredibly limited,” Altman writes, but “good enough at some things to create a misleading impression of greatness.”
Currently, he says, it’s a mistake to use ChatGPT for important tasks. The system is a glimpse of progress; in terms of robustness and reliability, there is still much work to be done, Altman writes.
https://openai.com/api/policies/sharing-publication/
Also have gotten Stable Diffusion txt2img running
https://stability.ai
The instructions to install and community are on the Discord https://discord.gg/stablediffusion
There is a special install if you are running a GPU with less than 6GB VRAM by Basujindal.
txt2img is not very good at new and different really creative ideas and it can't make modifications, mainly new images. I thinks it's nVidia that are working on one which enables you to mask and regenerate. Not that Ive used it much.
Hope it can do good textures?
txt2txt with generating code, at this stage still requires massaging.
Some txt2txt chat prompts:
write a [length] [writing results] on [topic] in the style of [style]
Rephrase the text “”
Write me an outline of [book]
Elaborate on [section]
Replace my simple vocabulary with advanced english “”
Now write it in the style of [author]
Act as a professional spelling and grammar corrector and improver
now write that in the style of a pulitzer price winner
Act as an professional [insert profession]