search

Tencent / Hunyuan3D-1

Tencent Hunyuan3D-1.0: A Unified Framework for Text-to-3D and Image-to-3D Generation
https://3d.hunyuan.tencent.com/
Other
2.26k stars 169 forks source link

Update main.py #29

Open mst-rajatmishra opened 2 weeks ago

mst-rajatmishra commented 2 weeks ago

Hi, I tried making some improvments, Added --verbose, --save_intermediate, --output_name to improve usability and customization. Arguments like --do_texture_mapping and --save_memory were preserved.

The check_paths() function ensures that all the model files and configurations are valid before running the pipeline. Descriptive error messages are shown if files or paths are missing.

tqdm could be integrated for tasks that have loops, but here it's mainly for checking and saving intermediate outputs. Added --verbose flag for users to get step-by-step feedback in the terminal.

The output filenames are now dynamically named using the --output_name argument, allowing users to customize their output filenames for mesh files and gifs. Intermediate images are saved in images subfolder, meshes in models, and gifs in renders.

Automatically creates necessary subfolders (images, models, renders) inside the output folder for better file organization.

The --verbose flag provides additional print statements, giving users progress updates. Intermediate Results:

The pipeline now saves intermediate steps (images and views) if --save_intermediate is specified. Thank You, RAJAT MISHRA

skypeandyk7 commented 2 weeks ago

Make it greater and greater!