NERF-SUPERVISION 生成新数据集

1、 https://github.com/yenchenlin/nerf-supervision-public

2、conda

3、问题

1、ModuleNotFoundError: No module named 'skimage'

pip install scikit-image

2、TypeError: read() got an unexpected keyword argument 'ignoregamma'

NeRF运行遇到的两个错误 - 知乎

return imageio.imread(f, format="PNG-PIL", ignoregamma=True)

3、RuntimeError: Expected a 'cuda' device type for generator but found 'cpu'

https://github.com/dbolya/yolact/issues/664

#raysRGB_iter = iter(DataLoader(RayDataset(rays_rgb), batch_size = N_rand, shuffle=True, num_workers=0))

raysRGB_iter = iter(DataLoader(RayDataset(rays_rgb), batch_size = N_rand, shuffle=False, num_workers=0))

shuffle:(数据类型 bool)

洗牌。默认设置为False。在每次迭代训练时是否将数据洗牌,默认设置是False。将输入数据的顺序打乱,是为了使数据更有独立性,但如果数据是有序列特征的,就不要设置成True了

猜你喜欢

转载自blog.csdn.net/zenglongjian/article/details/131022939