[Feature] RuntimeError: Tensors of type TensorImpl do not have sizes

[chfyyds]

What is the feature?

Please tell me why I encountered such a problem when training bevfusion. Hope to get your answer.

Traceback (most recent call last): File "/media/chf/新加卷/mmdetection3d/tools/train.py", line 149, in main() File "/media/chf/新加卷/mmdetection3d/tools/train.py", line 145, in main runner.train() File "/media/chf/25ee13e8-f20b-4792-9213-0157bc0af688/chf/anaconda3/envs/mmdetection3d/lib/python3.8/site-packages/mmengine/runner/runner.py", line 1745, in train model = self.train_loop.run() # type: ignore File "/media/chf/25ee13e8-f20b-4792-9213-0157bc0af688/chf/anaconda3/envs/mmdetection3d/lib/python3.8/site-packages/mmengine/runner/loops.py", line 96, in run self.run_epoch() File "/media/chf/25ee13e8-f20b-4792-9213-0157bc0af688/chf/anaconda3/envs/mmdetection3d/lib/python3.8/site-packages/mmengine/runner/loops.py", line 112, in run_epoch self.run_iter(idx, data_batch) File "/media/chf/25ee13e8-f20b-4792-9213-0157bc0af688/chf/anaconda3/envs/mmdetection3d/lib/python3.8/site-packages/mmengine/runner/loops.py", line 128, in run_iter outputs = self.runner.model.train_step( File "/media/chf/25ee13e8-f20b-4792-9213-0157bc0af688/chf/anaconda3/envs/mmdetection3d/lib/python3.8/site-packages/mmengine/model/base_model/base_model.py", line 114, in train_step losses = self._run_forward(data, mode='loss') # type: ignore File "/media/chf/25ee13e8-f20b-4792-9213-0157bc0af688/chf/anaconda3/envs/mmdetection3d/lib/python3.8/site-packages/mmengine/model/base_model/base_model.py", line 340, in _run_forward results = self(data, mode=mode) File "/media/chf/25ee13e8-f20b-4792-9213-0157bc0af688/chf/anaconda3/envs/mmdetection3d/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl return forward_call(input, kwargs) File "/media/chf/新加卷/mmdetection3d/mmdet3d/models/detectors/base.py", line 75, in forward return self.loss(inputs, data_samples, kwargs) File "/media/chf/新加卷/mmdetection3d/projects/BEVFusion/bevfusion/bevfusion.py", line 290, in loss feats = self.extract_feat(batch_inputs_dict, batch_input_metas) File "/media/chf/新加卷/mmdetection3d/projects/BEVFusion/bevfusion/bevfusion.py", line 266, in extract_feat img_feature = self.extract_img_feat(imgs, deepcopy(points), File "/media/chf/新加卷/mmdetection3d/projects/BEVFusion/bevfusion/bevfusion.py", line 154, in extract_img_feat x = self.view_transform( File "/media/chf/25ee13e8-f20b-4792-9213-0157bc0af688/chf/anaconda3/envs/mmdetection3d/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl return forward_call(input, kwargs) File "/media/chf/新加卷/mmdetection3d/projects/BEVFusion/bevfusion/depth_lss.py", line 424, in forward x = super().forward(*args, **kwargs) File "/media/chf/新加卷/mmdetection3d/projects/BEVFusion/bevfusion/depth_lss.py", line 330, in forward x = self.bev_pool(geom, x) File "/media/chf/新加卷/mmdetection3d/projects/BEVFusion/bevfusion/depth_lss.py", line 143, in bev_pool x = bev_pool(x, geom_feats, B, self.nx[2], self.nx[0], self.nx[1]) File "/media/chf/新加卷/mmdetection3d/projects/BEVFusion/bevfusion/ops/bev_pool/bev_pool.py", line 92, in bev_pool x = QuickCumsumCuda.apply(feats, coords, ranks, B, D, H, W) File "/media/chf/新加卷/mmdetection3d/projects/BEVFusion/bevfusion/ops/bev_pool/bev_pool.py", line 48, in forward out = bev_pool_ext.bev_pool_forward( RuntimeError: Tensors of type TensorImpl do not have sizes

Any other context?

No response

[2460707925]

hello , do you solve this

[wrencanfly]

Reinstall package and you will be good I guess