RuntimeError: mat1 and mat2 shapes cannot be multiplied (0x1 and 64x128) #56
Comments
packages in environment at D:\Gaussian\GaussianTalker\venv:Name Version Build Channelabsl-py 2.1.0 pypi_0 pypi |
|
hello. I have some issues I would like to ask you about setting up an environment for this project. How can I contact you? |
|
I added a linear layer to align the two matrices. |
(D:\Gaussian\GaussianTalker\venv) D:\Gaussian\GaussianTalker>python train.py -s .\data\obama --model_path .\data\obama\models --configs arguments\64_dim_1_transformer.py
Optimizing .\data\obama\models
Output folder: .\data\obama\models [15/09 23:22:08]
feature_dim: 64 [15/09 23:22:08]
[INFO] load aud_features: torch.Size([7999, 29, 16]) [15/09 23:22:08]
Reading Training Transforms [15/09 23:22:09]
Reading Test Transforms [15/09 23:24:03]
Generating Video Transforms [15/09 23:24:18]
Loading Training Cameras [15/09 23:24:26]
Loading Test Cameras [15/09 23:24:26]
Loading Video Cameras [15/09 23:24:26]
Deformation Net Set aabb [0.75368124 0.8865968 0.49712884] [-0.7633047 -0.859515 -0.5761221] [15/09 23:24:26]
Voxel Plane: set aabb= Parameter containing:
tensor([[ 0.7537, 0.8866, 0.4971],
[-0.7633, -0.8595, -0.5761]]) [15/09 23:24:26]
Number of points at initialisation : 34650 [15/09 23:24:26]
['xyz', 'deformation', 'grid', 'f_dc', 'f_rest', 'opacity', 'scaling', 'rotation'] [15/09 23:24:26]
Training progress: 0%| | 0/7999 [00:00<?, ?it/s]data loading done [15/09 23:24:26]
[ITER 1] Saving Gaussians [15/09 23:35:54]
Training progress: 6%|█▏ | 500/7999 [12:10<10:14, 12.20it/s, Loss=0.2446528, psnr=20.31, point=34650]
[ITER 500] Saving Gaussians [15/09 23:36:36]
Training progress: 11%|██▌ | 900/7999 [12:41<08:50, 13.38it/s, Loss=0.2448057, psnr=20.14, point=2]Traceback (most recent call last):
File "D:\Gaussian\GaussianTalker\train.py", line 408, in
training(lp.extract(args), hp.extract(args), op.extract(args), pp.extract(args), args.test_iterations, args.save_iterations, args.checkpoint_iterations, args.start_checkpoint, args.debug_from, args.expname, args.use_wandb)
File "D:\Gaussian\GaussianTalker\train.py", line 276, in training
scene_reconstruction(dataset, opt, hyper, pipe, testing_iterations, saving_iterations,
File "D:\Gaussian\GaussianTalker\train.py", line 153, in scene_reconstruction
output = render_from_batch(viewpoint_cams, gaussians, pipe, random_color, stage=stage, batch_size=batch_size, canonical_tri_plane_factor_list=opt.canonical_tri_plane_factor_list,iteration=iteration)
File "D:\Gaussian\GaussianTalker\gaussian_renderer_init_.py", line 101, in render_from_batch
means3D_final, scales_temp, rotations_temp, opacity_temp, shs_temp = pc._deformation(means3D, scales, rotations, opacity, shs, aud_features, eye_features, cam_features)
File "D:\Gaussian\GaussianTalker\venv\lib\site-packages\torch\nn\modules\module.py", line 1518, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "D:\Gaussian\GaussianTalker\venv\lib\site-packages\torch\nn\modules\module.py", line 1527, in _call_impl
return forward_call(*args, **kwargs)
File "D:\Gaussian\GaussianTalker\scene\deformation.py", line 365, in forward
return self.forward_static(point, scales, rotations, opacity, shs)
File "D:\Gaussian\GaussianTalker\scene\deformation.py", line 375, in forward_static
points = self.deformation_net(points)
File "D:\Gaussian\GaussianTalker\venv\lib\site-packages\torch\nn\modules\module.py", line 1518, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "D:\Gaussian\GaussianTalker\venv\lib\site-packages\torch\nn\modules\module.py", line 1527, in _call_impl
return forward_call(*args, **kwargs)
File "D:\Gaussian\GaussianTalker\scene\deformation.py", line 211, in forward
return self.forward_static(rays_pts_emb)
File "D:\Gaussian\GaussianTalker\scene\deformation.py", line 218, in forward_static
grid_feature, scale, rotation, opacity, sh = self.tri_plane(rays_pts_emb)
File "D:\Gaussian\GaussianTalker\venv\lib\site-packages\torch\nn\modules\module.py", line 1518, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "D:\Gaussian\GaussianTalker\venv\lib\site-packages\torch\nn\modules\module.py", line 1527, in _call_impl
return forward_call(*args, **kwargs)
File "D:\Gaussian\GaussianTalker\scene\canonical_tri_plane.py", line 54, in forward
feature = self.feature_out(feature)
File "D:\Gaussian\GaussianTalker\venv\lib\site-packages\torch\nn\modules\module.py", line 1518, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "D:\Gaussian\GaussianTalker\venv\lib\site-packages\torch\nn\modules\module.py", line 1527, in _call_impl
return forward_call(*args, **kwargs)
File "D:\Gaussian\GaussianTalker\venv\lib\site-packages\torch\nn\modules\container.py", line 215, in forward
input = module(input)
File "D:\Gaussian\GaussianTalker\venv\lib\site-packages\torch\nn\modules\module.py", line 1518, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "D:\Gaussian\GaussianTalker\venv\lib\site-packages\torch\nn\modules\module.py", line 1527, in _call_impl
return forward_call(*args, **kwargs)
File "D:\Gaussian\GaussianTalker\venv\lib\site-packages\torch\nn\modules\linear.py", line 114, in forward
return F.linear(input, self.weight, self.bias)
RuntimeError: mat1 and mat2 shapes cannot be multiplied (0x1 and 64x128)
Training progress: 11%|██▎ | 900/7999 [12:44<1:40:26, 1.18it/s, Loss=0.2448057, psnr=20.14, point=2]
The text was updated successfully, but these errors were encountered: