search

nianticlabs / diffusionerf

[CVPR 2023] DiffusioNeRF: Regularizing Neural Radiance Fields with Denoising Diffusion Models
MIT License
296 stars 17 forks source link

train error #21

Open XLR-man opened 7 months ago

XLR-man commented 7 months ago

i download pretrained diffusion model you provided，but when i train，it got error： ![Uploading error.png…]()

  File "/data/xielangren/project/Aleth-NeRF/src/model/llnerf/model.py", line 81, in setup
    patch_diffusion_model = load_patch_diffusion_model(Path(self.patch_regulariser_path))
  File "/data/xielangren/project/Aleth-NeRF/src/model/llnerf/learned_regularisation/patch_regulariser.py", line 138, in load_patch_diffusion_model
    trainer.load(str(reg_checkpoint_path))
  File "/data/xielangren/project/Aleth-NeRF/src/model/llnerf/learned_regularisation/diffusion/denoising_diffusion_pytorch.py", line 836, in load
    self.ema.load_state_dict(data['ema'])
AttributeError: 'Trainer' object has no attribute 'ema'

how to fix the error

XLR-man commented 7 months ago

I found that this seems to be the case for multi-GPU training, but not for single-GPU training. Why is that?