None of the files define DiTGenerator or a generator class with a name matching the config's expectation.
You have classes like Autoencoder, Encoder, Decoder, BGModel, Raymarcher, but not a top-level 3D diffusion generator.
What Does This Mean?
The official repo does not include the generator model code needed to run inference as designed.
The config and pipeline expect a class that is missing (DiTGenerator or similar).
The available classes are for encoders, decoders, background models, etc.—not for standalone text-to-3D generation.
Summary
The required model class (DiTGenerator) is missing.
Hi, thanks for this repo, but I'm facing some issues runing it on www.kaggle.com
I got a quote above, missing a summary
I have this notebook currently...
3dtopia-xl-kaggle-full_Version2.ipynb
Please take a look at
8. config file
import yaml
config = {
'output_dir': './outputs/text_hq',
'inference': {'ddim': 90, 'cfg': 7.0, 'seed': 1234},
'model': {
'generator': {
'class_name': 'dva.models.dit.DiTGenerator',
is this path dva.models.dit.DiTGenerator, it's ok?
Do you know where I can find it?
'generator': {
'class_name': 'dva.models.dit.DiTGenerator', # Edit path if needed after checking your repo!
Thanks in advance
Best regards
Carlos Vieira