作者你好！我在做遥感任务会涉及到多通道，我发现__init__.py中似乎没有涉及到对于多通道的实现？渲染报错如下
Traceback (most recent call last):
File "/home/lizhuomao/Contrast/restruct/MSDDN-main/main.py", line 20, in
solver.run()
File "/home/lizhuomao/Contrast/restruct/MSDDN-main/solver/solver.py", line 442, in run
self.train()
File "/home/lizhuomao/Contrast/restruct/MSDDN-main/solver/solver.py", line 234, in train
sr = self.model(lms, pan, render_settings)
File "/home/lizhuomao/anaconda3/envs/gsplat_cu118/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "/home/lizhuomao/anaconda3/envs/gsplat_cu118/lib/python3.9/site-packages/torch/nn/parallel/data_parallel.py", line 169, in forward
return self.module(*inputs[0], **kwargs[0])
File "/home/lizhuomao/anaconda3/envs/gsplat_cu118/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "/home/lizhuomao/Contrast/restruct/MSDDN-main/model/gspan.py", line 277, in forward
rendered_hrms_image = render_gaussians(gaussians_dataclass, render_set)
File "/home/lizhuomao/Contrast/restruct/MSDDN-main/model/gspan.py", line 71, in render_gaussians
img = _gaussian2image(mean, scale, rot, feature, render_set, opacity)
File "/home/lizhuomao/Contrast/restruct/MSDDN-main/model/gspan.py", line 55, in _gaussian2image
out_img = rasterize_gaussians_sum(
File "/home/lizhuomao/gsplat-main/gsplat/rasterize_sum.py", line 72, in rasterize_gaussians_sum
return _RasterizeGaussiansSum.apply(
File "/home/lizhuomao/anaconda3/envs/gsplat_cu118/lib/python3.9/site-packages/torch/autograd/function.py", line 506, in apply
return super().apply(*args, **kwargs) # type: ignore[misc]
File "/home/lizhuomao/gsplat-main/gsplat/rasterize_sum.py", line 149, in forward
rasterize_fn = _C.nd_rasterize_sum_forward
AttributeError: module 'gsplat.cuda' has no attribute 'nd_rasterize_sum_forward'
所以是否缺少
nd_rasterize_sum_forward = make_lazy_cuda_func("nd_rasterize_sum_forward")
nd_rasterize_sum_backward = make_lazy_cuda_func("nd_rasterize_sum_backward")等代码？
我需要修改__init.py后再重新安装gsplat吗？我对这个不太熟悉。谢谢！
作者你好！我在做遥感任务会涉及到多通道，我发现__init__.py中似乎没有涉及到对于多通道的实现？渲染报错如下
Traceback (most recent call last):
File "/home/lizhuomao/Contrast/restruct/MSDDN-main/main.py", line 20, in
solver.run()
File "/home/lizhuomao/Contrast/restruct/MSDDN-main/solver/solver.py", line 442, in run
self.train()
File "/home/lizhuomao/Contrast/restruct/MSDDN-main/solver/solver.py", line 234, in train
sr = self.model(lms, pan, render_settings)
File "/home/lizhuomao/anaconda3/envs/gsplat_cu118/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "/home/lizhuomao/anaconda3/envs/gsplat_cu118/lib/python3.9/site-packages/torch/nn/parallel/data_parallel.py", line 169, in forward
return self.module(*inputs[0], **kwargs[0])
File "/home/lizhuomao/anaconda3/envs/gsplat_cu118/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "/home/lizhuomao/Contrast/restruct/MSDDN-main/model/gspan.py", line 277, in forward
rendered_hrms_image = render_gaussians(gaussians_dataclass, render_set)
File "/home/lizhuomao/Contrast/restruct/MSDDN-main/model/gspan.py", line 71, in render_gaussians
img = _gaussian2image(mean, scale, rot, feature, render_set, opacity)
File "/home/lizhuomao/Contrast/restruct/MSDDN-main/model/gspan.py", line 55, in _gaussian2image
out_img = rasterize_gaussians_sum(
File "/home/lizhuomao/gsplat-main/gsplat/rasterize_sum.py", line 72, in rasterize_gaussians_sum
return _RasterizeGaussiansSum.apply(
File "/home/lizhuomao/anaconda3/envs/gsplat_cu118/lib/python3.9/site-packages/torch/autograd/function.py", line 506, in apply
return super().apply(*args, **kwargs) # type: ignore[misc]
File "/home/lizhuomao/gsplat-main/gsplat/rasterize_sum.py", line 149, in forward
rasterize_fn = _C.nd_rasterize_sum_forward
AttributeError: module 'gsplat.cuda' has no attribute 'nd_rasterize_sum_forward'
所以是否缺少
nd_rasterize_sum_forward = make_lazy_cuda_func("nd_rasterize_sum_forward")
nd_rasterize_sum_backward = make_lazy_cuda_func("nd_rasterize_sum_backward")等代码？
我需要修改__init.py后再重新安装gsplat吗？我对这个不太熟悉。谢谢！