SWUSTLKL
commented
6 months ago May I ask why the GPU is so large because of the model or the OBJ I use?| I'm running the piece of code that generated the dataset。 python import_lenses.py --input_path path_to_target
Hello, while running your code, I used a GeForce RTX 3080 with an obj file of size 3.74MB and got the error like this:
cuda is available. Traceback (most recent call last): File "E:\allpaper\Camera-calibration\Neural Lens Modeling\code\NeuroLens\SynLens\import_lenses.py", line 111, in
process_mesh(lenses_data[i], lens_name, num_views, image_size, render_rgb, out_fd)
File "E:\allpaper\Camera-calibration\Neural Lens Modeling\code\NeuroLens\SynLens\renderer.py", line 413, in process_mesh
target_cameras, target_rgb = render_mesh(obj_filename, num_views, image_size, render_rgb, lens_model)
File "E:\allpaper\Camera-calibration\Neural Lens Modeling\code\NeuroLens\SynLens\renderer.py", line 393, in render_mesh
target_images = renderer(meshes, cameras=cameras, lights=lights)
File "D:\anaconda\envs\neural\lib\site-packages\torch\nn\modules\module.py", line 1518, in _wrapped_call_impl
return self._call_impl(*args, kwargs)
File "D:\anaconda\envs\neural\lib\site-packages\torch\nn\modules\module.py", line 1527, in _call_impl
return forward_call(args, kwargs)
File "D:\anaconda\envs\neural\lib\site-packages\pytorch3d-0.7.5-py3.9-win-amd64.egg\pytorch3d\renderer\mesh\renderer.py", line 62, in forward
images = self.shader(fragments, meshes_world, kwargs)
File "D:\anaconda\envs\neural\lib\site-packages\torch\nn\modules\module.py", line 1518, in _wrapped_call_impl
return self._call_impl(args, kwargs)
File "D:\anaconda\envs\neural\lib\site-packages\torch\nn\modules\module.py", line 1527, in _call_impl
return forward_call(*args, **kwargs)
File "D:\anaconda\envs\neural\lib\site-packages\pytorch3d-0.7.5-py3.9-win-amd64.egg\pytorch3d\renderer\mesh\shader.py", line 126, in forward
texels = meshes.sample_textures(fragments)
File "D:\anaconda\envs\neural\lib\site-packages\pytorch3d-0.7.5-py3.9-win-amd64.egg\pytorch3d\structures\meshes.py", line 1542, in sample_textures
return self.textures.sample_textures(
File "D:\anaconda\envs\neural\lib\site-packages\pytorch3d-0.7.5-py3.9-win-amd64.egg\pytorch3d\renderer\mesh\textures.py", line 1003, in sample_textures
texels = F.grid_sample(
File "D:\anaconda\envs\neural\lib\site-packages\torch\nn\functional.py", line 4304, in grid_sample
return torch.grid_sampler(input, grid, mode_enum, padding_mode_enum, align_corners)
torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 55.57 GiB. GPU 0 has a total capacty of 12.00 GiB of which 9.02 GiB is free. Of the allocated memory 984.52 MiB is allocated by PyTorch, and 463.48 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF
Is it possible that I need more than 55.57 GB of video memory because I chose an inappropriate obj file? Is there a recommended demo program suitable for 12g video memory?