Update mesh export in inference.py to use .glb suffix for temporary files and return the correct path without appending .obj. This improves file handling consistency.

Refactor mesh export process in inference.py to directly save as GLB format, removing intermediate OBJ conversion steps. Update return value to reflect new output format.

Refactor generate3d function in inference.py to enhance readability by removing unnecessary blank lines. Update requirements.txt to specify versions for gradio and huggingface-hub for better dependency management.

Refactor generate3d function in inference.py to improve readability by removing unnecessary blank lines. Update requirements.txt to remove gradio and numpy versions for cleaner dependency management.

Update inference.py to output glb on generate 3d

remove cuda

float32

remove codes to inference

init