POINT-E is a neural network that creates a cloud of points from text or an image that can be translated into a 3D Model (Text / Image-> 3D) created by OpenAI
- Python 3.10.6
- Internet connection (POINT-E loads from the Internet models that are not in the folder, if there is, then it starts loading into memory and generating)
- Git
- Install pytorch from official website for CPU or GPU
- Run in the main folder:
pip install -r requirements.txt
Run the main.py file
Example: python main.py
Q: How to update this WebUI?
A: You need run in main webui folder: git pull
Q: How to update POINT-E?
A: Just like WebUI, only you have to go to the point-e folder and run git pull
Q: If I have an error ModuleNotFoundError, what I need to do?
A: Try run pip install -r requirements.txt again. The fact is that if a new version of the script requires some kind of library, you must install it.
Q: Where are my results saved?
A: In the outputs/ folder there are two folders, one model created by text and the other by images.
Q: What are the files inside the outputs folder?
A: The first file is a *.ply file with a point cloud The second file is a *.ply file with a mesh The third file is a *.obj file, the same with a mesh only converted from *.ply
Q: What is the point_e_modelcache folder?
A: This is the folder where the POINT-E models are located, which are downloaded to generate 3d models.