search

open-mmlab / mmdetection3d

OpenMMLab's next-generation platform for general 3D object detection.
https://mmdetection3d.readthedocs.io/en/latest/
Apache License 2.0
5.3k stars 1.54k forks source link

Why does this demo have mistake? #2474

Open xiaomonv111 opened 1 year ago

xiaomonv111 commented 1 year ago

from mmdet3d.apis import inference_detector, init_model from mmdet3d.registry import VISUALIZERS from mmdet3d.utils import register_all_modules config_file = '/home/dxy/deepl/mmdetection3d/configs/second/second_hv_secfpn_8xb6-80e_kitti-3d-3class.py' checkpoint='/home/dxy/deepl/mmdetection3d/checkpoints/hv_second_secfpn_6x8_80e_kitti-3d-3class_20210831_022017-ae782e87.pth' model = init_model(config_file, checkpoint, device='cuda:0') visualizer = VISUALIZERS.build(model.cfg.visualizer) visualizer = VISUALIZERS.build(model.cfg.visualizer) visualizer.dataset_meta = { 'CLASSES': model.CLASSES, 'PALETTE': model.PALETTE }

mistake：

AttributeError Traceback (most recent call last) /tmp/ipykernel_23136/1014200392.py in 2 visualizer = VISUALIZERS.build(model.cfg.visualizer) 3 visualizer.dataset_meta = { ----> 4 'CLASSES': model.CLASSES, 5 'PALETTE': model.PALETTE 6 } ~/miniconda3/envs/openmmlab/lib/python3.7/site-packages/torch/nn/modules/module.py in getattr(self, name) 1268 return modules[name] 1269 raise AttributeError("'{}' object has no attribute '{}'".format( -> 1270 type(self).name, name)) 1271 1272 def setattr(self, name: str, value: Union[Tensor, 'Module']) -> None:

AttributeError: 'VoxelNet' object has no attribute 'CLASSES'

Why does this demo have mistake? I choose the second. Thank you.

JingweiZhang12 commented 1 year ago

Suggest using the dev-1.x branch and it's the latest code. The demo in dev-1.x branch has been verified.

xiaomonv111 commented 1 year ago

Sorry, I feel that there is still a problem with dev1.x. It is the same problem. Is there any other correction plan? Can I ask if checkpoint_file is to download hv_second_secfpn_6x8_80e_kitti-3d-3class_20210831_022017-ae782e87.pth' on github

JingweiZhang12 commented 1 year ago

  1. The checkpoint link of SECOND has been updated in https://github.com/open-mmlab/mmdetection3d/blob/dev-1.x/configs/second/README.md

  2. The PALETTE and CLASSES should be in lower case format and these keys are under model.dataset_meta. https://github.com/open-mmlab/mmdetection3d/blob/dev-1.x/mmdet3d/apis/inference.py#L82

xiaomonv111 commented 1 year ago

very sorry,it still have mistake.

init visualizer

visualizer = VISUALIZERS.build(model.cfg.visualizer) visualizer.dataset_meta = { 'classes': model.classes, 'palette': model.palette }

AttributeError Traceback (most recent call last) /tmp/ipykernel_23136/390537073.py in 2 visualizer = VISUALIZERS.build(model.cfg.visualizer) 3 visualizer.dataset_meta = { ----> 4 'classes': model.classes, 5 'palette': model.palette 6 }

~/miniconda3/envs/openmmlab/lib/python3.7/site-packages/torch/nn/modules/module.py in getattr(self, name) 1268 return modules[name] 1269 raise AttributeError("'{}' object has no attribute '{}'".format( -> 1270 type(self).name, name)) 1271 1272 def setattr(self, name: str, value: Union[Tensor, 'Module']) -> None:

AttributeError: 'VoxelNet' object has no attribute 'classes'

Very thank you

JingweiZhang12 commented 1 year ago

@xiaomonv111 Hi, you could refer to this demo example to customize your demo.

XinyueZ commented 1 year ago

Run at dev-1.x in inference_demo.ipynb @JingweiZhang12 

# test a single sample
pcd = './data/kitti/000008.bin'
result, data = inference_detector(model, pcd)
points = data['inputs']['points']
data_input = dict(points=points)

Error:

--------------------------------------------------------------------------
RuntimeError                              Traceback (most recent call last)
[/workspaces/smartGate/trainers/mmdetection3d/demo/inference_demo.ipynb](https://vscode-remote+dev-002dcontainer-002b7b22686f737450617468223a222f686f6d652f7562756e74752f776f726b2f736d61727447617465222c2273657474696e6773223a7b22686f7374223a227373683a2f2f3139322e3136382e332e313737227d7d.vscode-resource.vscode-cdn.net/workspaces/smartGate/trainers/mmdetection3d/demo/inference_demo.ipynb) Cell 8 line 3
      [1](vscode-notebook-cell://dev-container%2B7b22686f737450617468223a222f686f6d652f7562756e74752f776f726b2f736d61727447617465222c2273657474696e6773223a7b22686f7374223a227373683a2f2f3139322e3136382e332e313737227d7d/workspaces/smartGate/trainers/mmdetection3d/demo/inference_demo.ipynb#X24sdnNjb2RlLXJlbW90ZQ%3D%3D?line=0) # test a single sample
      [2](vscode-notebook-cell://dev-container%2B7b22686f737450617468223a222f686f6d652f7562756e74752f776f726b2f736d61727447617465222c2273657474696e6773223a7b22686f7374223a227373683a2f2f3139322e3136382e332e313737227d7d/workspaces/smartGate/trainers/mmdetection3d/demo/inference_demo.ipynb#X24sdnNjb2RlLXJlbW90ZQ%3D%3D?line=1) pcd = '[./data/kitti/000008.bin](https://vscode-remote+dev-002dcontainer-002b7b22686f737450617468223a222f686f6d652f7562756e74752f776f726b2f736d61727447617465222c2273657474696e6773223a7b22686f7374223a227373683a2f2f3139322e3136382e332e313737227d7d.vscode-resource.vscode-cdn.net/workspaces/smartGate/trainers/mmdetection3d/demo/data/kitti/000008.bin)'
----> [3](vscode-notebook-cell://dev-container%2B7b22686f737450617468223a222f686f6d652f7562756e74752f776f726b2f736d61727447617465222c2273657474696e6773223a7b22686f7374223a227373683a2f2f3139322e3136382e332e313737227d7d/workspaces/smartGate/trainers/mmdetection3d/demo/inference_demo.ipynb#X24sdnNjb2RlLXJlbW90ZQ%3D%3D?line=2) result, data = inference_detector(model, pcd)
      [4](vscode-notebook-cell://dev-container%2B7b22686f737450617468223a222f686f6d652f7562756e74752f776f726b2f736d61727447617465222c2273657474696e6773223a7b22686f7374223a227373683a2f2f3139322e3136382e332e313737227d7d/workspaces/smartGate/trainers/mmdetection3d/demo/inference_demo.ipynb#X24sdnNjb2RlLXJlbW90ZQ%3D%3D?line=3) points = data['inputs']['points']
      [5](vscode-notebook-cell://dev-container%2B7b22686f737450617468223a222f686f6d652f7562756e74752f776f726b2f736d61727447617465222c2273657474696e6773223a7b22686f7374223a227373683a2f2f3139322e3136382e332e313737227d7d/workspaces/smartGate/trainers/mmdetection3d/demo/inference_demo.ipynb#X24sdnNjb2RlLXJlbW90ZQ%3D%3D?line=4) data_input = dict(points=points)

File [~/.local/lib/python3.10/site-packages/mmdet3d/apis/inference.py:182](https://vscode-remote+dev-002dcontainer-002b7b22686f737450617468223a222f686f6d652f7562756e74752f776f726b2f736d61727447617465222c2273657474696e6773223a7b22686f7374223a227373683a2f2f3139322e3136382e332e313737227d7d.vscode-resource.vscode-cdn.net/workspaces/smartGate/trainers/mmdetection3d/demo/~/.local/lib/python3.10/site-packages/mmdet3d/apis/inference.py:182), in inference_detector(model, pcds)
    180 # forward the model
    181 with torch.no_grad():
--> 182     results = model.test_step(collate_data)
    184 if not is_batch:
    185     return results[0], data[0]

File [/usr/local/lib/python3.10/dist-packages/mmengine/model/base_model/base_model.py:145](https://vscode-remote+dev-002dcontainer-002b7b22686f737450617468223a222f686f6d652f7562756e74752f776f726b2f736d61727447617465222c2273657474696e6773223a7b22686f7374223a227373683a2f2f3139322e3136382e332e313737227d7d.vscode-resource.vscode-cdn.net/usr/local/lib/python3.10/dist-packages/mmengine/model/base_model/base_model.py:145), in BaseModel.test_step(self, data)
    136 """``BaseModel`` implements ``test_step`` the same as ``val_step``.
    137 
    138 Args:
   (...)
    142     list: The predictions of given data.
    143 """
    144 data = self.data_preprocessor(data, False)
--> 145 return self._run_forward(data, mode='predict')

File [/usr/local/lib/python3.10/dist-packages/mmengine/model/base_model/base_model.py:340](https://vscode-remote+dev-002dcontainer-002b7b22686f737450617468223a222f686f6d652f7562756e74752f776f726b2f736d61727447617465222c2273657474696e6773223a7b22686f7374223a227373683a2f2f3139322e3136382e332e313737227d7d.vscode-resource.vscode-cdn.net/usr/local/lib/python3.10/dist-packages/mmengine/model/base_model/base_model.py:340), in BaseModel._run_forward(self, data, mode)
    330 """Unpacks data for :meth:`forward`
    331 
    332 Args:
   (...)
    337     dict or list: Results of training or testing mode.
    338 """
    339 if isinstance(data, dict):
--> 340     results = self(**data, mode=mode)
    341 elif isinstance(data, (list, tuple)):
    342     results = self(*data, mode=mode)

File [/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py:1501](https://vscode-remote+dev-002dcontainer-002b7b22686f737450617468223a222f686f6d652f7562756e74752f776f726b2f736d61727447617465222c2273657474696e6773223a7b22686f7374223a227373683a2f2f3139322e3136382e332e313737227d7d.vscode-resource.vscode-cdn.net/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py:1501), in Module._call_impl(self, *args, **kwargs)
   1496 # If we don't have any hooks, we want to skip the rest of the logic in
   1497 # this function, and just call forward.
   1498 if not (self._backward_hooks or self._backward_pre_hooks or self._forward_hooks or self._forward_pre_hooks
   1499         or _global_backward_pre_hooks or _global_backward_hooks
   1500         or _global_forward_hooks or _global_forward_pre_hooks):
-> 1501     return forward_call(*args, **kwargs)
   1502 # Do not call functions when jit is used
   1503 full_backward_hooks, non_full_backward_hooks = [], []

File [~/.local/lib/python3.10/site-packages/mmdet3d/models/detectors/base.py:86](https://vscode-remote+dev-002dcontainer-002b7b22686f737450617468223a222f686f6d652f7562756e74752f776f726b2f736d61727447617465222c2273657474696e6773223a7b22686f7374223a227373683a2f2f3139322e3136382e332e313737227d7d.vscode-resource.vscode-cdn.net/workspaces/smartGate/trainers/mmdetection3d/demo/~/.local/lib/python3.10/site-packages/mmdet3d/models/detectors/base.py:86), in Base3DDetector.forward(self, inputs, data_samples, mode, **kwargs)
     84         return self.aug_test(inputs, data_samples, **kwargs)
     85     else:
---> 86         return self.predict(inputs, data_samples, **kwargs)
     87 elif mode == 'tensor':
     88     return self._forward(inputs, data_samples, **kwargs)

File [~/.local/lib/python3.10/site-packages/mmdet3d/models/detectors/single_stage.py:109](https://vscode-remote+dev-002dcontainer-002b7b22686f737450617468223a222f686f6d652f7562756e74752f776f726b2f736d61727447617465222c2273657474696e6773223a7b22686f7374223a227373683a2f2f3139322e3136382e332e313737227d7d.vscode-resource.vscode-cdn.net/workspaces/smartGate/trainers/mmdetection3d/demo/~/.local/lib/python3.10/site-packages/mmdet3d/models/detectors/single_stage.py:109), in SingleStage3DDetector.predict(self, batch_inputs_dict, batch_data_samples, **kwargs)
     78 def predict(self, batch_inputs_dict: dict, batch_data_samples: SampleList,
     79             **kwargs) -> SampleList:
     80     """Predict results from a batch of inputs and data samples with post-
     81     processing.
     82 
   (...)
    107                 (num_instances, C) where C >=7.
    108     """
--> 109     x = self.extract_feat(batch_inputs_dict)
    110     results_list = self.bbox_head.predict(x, batch_data_samples, **kwargs)
    111     predictions = self.add_pred_to_datasample(batch_data_samples,
    112                                               results_list)

File [~/.local/lib/python3.10/site-packages/mmdet3d/models/detectors/voxelnet.py:43](https://vscode-remote+dev-002dcontainer-002b7b22686f737450617468223a222f686f6d652f7562756e74752f776f726b2f736d61727447617465222c2273657474696e6773223a7b22686f7374223a227373683a2f2f3139322e3136382e332e313737227d7d.vscode-resource.vscode-cdn.net/workspaces/smartGate/trainers/mmdetection3d/demo/~/.local/lib/python3.10/site-packages/mmdet3d/models/detectors/voxelnet.py:43), in VoxelNet.extract_feat(self, batch_inputs_dict)
     39 voxel_features = self.voxel_encoder(voxel_dict['voxels'],
     40                                     voxel_dict['num_points'],
     41                                     voxel_dict['coors'])
     42 batch_size = voxel_dict['coors'][-1, 0].item() + 1
---> 43 x = self.middle_encoder(voxel_features, voxel_dict['coors'],
     44                         batch_size)
     45 x = self.backbone(x)
     46 if self.with_neck:

File [/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py:1501](https://vscode-remote+dev-002dcontainer-002b7b22686f737450617468223a222f686f6d652f7562756e74752f776f726b2f736d61727447617465222c2273657474696e6773223a7b22686f7374223a227373683a2f2f3139322e3136382e332e313737227d7d.vscode-resource.vscode-cdn.net/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py:1501), in Module._call_impl(self, *args, **kwargs)
   1496 # If we don't have any hooks, we want to skip the rest of the logic in
   1497 # this function, and just call forward.
   1498 if not (self._backward_hooks or self._backward_pre_hooks or self._forward_hooks or self._forward_pre_hooks
   1499         or _global_backward_pre_hooks or _global_backward_hooks
   1500         or _global_forward_hooks or _global_forward_pre_hooks):
-> 1501     return forward_call(*args, **kwargs)
   1502 # Do not call functions when jit is used
   1503 full_backward_hooks, non_full_backward_hooks = [], []

File [/usr/lib/python3.10/contextlib.py:79](https://vscode-remote+dev-002dcontainer-002b7b22686f737450617468223a222f686f6d652f7562756e74752f776f726b2f736d61727447617465222c2273657474696e6773223a7b22686f7374223a227373683a2f2f3139322e3136382e332e313737227d7d.vscode-resource.vscode-cdn.net/usr/lib/python3.10/contextlib.py:79), in ContextDecorator.__call__.<locals>.inner(*args, **kwds)
     76 @wraps(func)
     77 def inner(*args, **kwds):
     78     with self._recreate_cm():
---> 79         return func(*args, **kwds)

File [~/.local/lib/python3.10/site-packages/mmdet3d/models/middle_encoders/sparse_encoder.py:145](https://vscode-remote+dev-002dcontainer-002b7b22686f737450617468223a222f686f6d652f7562756e74752f776f726b2f736d61727447617465222c2273657474696e6773223a7b22686f7374223a227373683a2f2f3139322e3136382e332e313737227d7d.vscode-resource.vscode-cdn.net/workspaces/smartGate/trainers/mmdetection3d/demo/~/.local/lib/python3.10/site-packages/mmdet3d/models/middle_encoders/sparse_encoder.py:145), in SparseEncoder.forward(self, voxel_features, coors, batch_size)
    142 coors = coors.int()
    143 input_sp_tensor = SparseConvTensor(voxel_features, coors,
    144                                    self.sparse_shape, batch_size)
--> 145 x = self.conv_input(input_sp_tensor)
    147 encode_features = []
    148 for encoder_layer in self.encoder_layers:

File [/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py:1501](https://vscode-remote+dev-002dcontainer-002b7b22686f737450617468223a222f686f6d652f7562756e74752f776f726b2f736d61727447617465222c2273657474696e6773223a7b22686f7374223a227373683a2f2f3139322e3136382e332e313737227d7d.vscode-resource.vscode-cdn.net/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py:1501), in Module._call_impl(self, *args, **kwargs)
   1496 # If we don't have any hooks, we want to skip the rest of the logic in
   1497 # this function, and just call forward.
   1498 if not (self._backward_hooks or self._backward_pre_hooks or self._forward_hooks or self._forward_pre_hooks
   1499         or _global_backward_pre_hooks or _global_backward_hooks
   1500         or _global_forward_hooks or _global_forward_pre_hooks):
-> 1501     return forward_call(*args, **kwargs)
   1502 # Do not call functions when jit is used
   1503 full_backward_hooks, non_full_backward_hooks = [], []

File [/usr/local/lib/python3.10/dist-packages/mmcv/ops/sparse_modules.py:135](https://vscode-remote+dev-002dcontainer-002b7b22686f737450617468223a222f686f6d652f7562756e74752f776f726b2f736d61727447617465222c2273657474696e6773223a7b22686f7374223a227373683a2f2f3139322e3136382e332e313737227d7d.vscode-resource.vscode-cdn.net/usr/local/lib/python3.10/dist-packages/mmcv/ops/sparse_modules.py:135), in SparseSequential.forward(self, input)
    133     assert isinstance(input, SparseConvTensor)
    134     self._sparity_dict[k] = input.sparity
--> 135     input = module(input)
    136 else:
    137     if isinstance(input, SparseConvTensor):

File [/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py:1501](https://vscode-remote+dev-002dcontainer-002b7b22686f737450617468223a222f686f6d652f7562756e74752f776f726b2f736d61727447617465222c2273657474696e6773223a7b22686f7374223a227373683a2f2f3139322e3136382e332e313737227d7d.vscode-resource.vscode-cdn.net/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py:1501), in Module._call_impl(self, *args, **kwargs)
   1496 # If we don't have any hooks, we want to skip the rest of the logic in
   1497 # this function, and just call forward.
   1498 if not (self._backward_hooks or self._backward_pre_hooks or self._forward_hooks or self._forward_pre_hooks
   1499         or _global_backward_pre_hooks or _global_backward_hooks
   1500         or _global_forward_hooks or _global_forward_pre_hooks):
-> 1501     return forward_call(*args, **kwargs)
   1502 # Do not call functions when jit is used
   1503 full_backward_hooks, non_full_backward_hooks = [], []

File [/usr/local/lib/python3.10/dist-packages/mmcv/ops/sparse_conv.py:157](https://vscode-remote+dev-002dcontainer-002b7b22686f737450617468223a222f686f6d652f7562756e74752f776f726b2f736d61727447617465222c2273657474696e6773223a7b22686f7374223a227373683a2f2f3139322e3136382e332e313737227d7d.vscode-resource.vscode-cdn.net/usr/local/lib/python3.10/dist-packages/mmcv/ops/sparse_conv.py:157), in SparseConvolution.forward(self, input)
    155         outids, _, indice_pairs, indice_pair_num, _ = data
    156     else:
--> 157         outids, indice_pairs, indice_pair_num = ops.get_indice_pairs(
    158             indices,
    159             batch_size,
    160             spatial_shape,
    161             self.kernel_size,
    162             self.stride,
    163             self.padding,
    164             self.dilation,
    165             self.output_padding,
    166             self.subm,
    167             self.transposed,
    168             grid=input.grid)
    169         input.indice_dict[self.indice_key] = (outids, indices,
    170                                               indice_pairs,
    171                                               indice_pair_num,
    172                                               spatial_shape)
    173 if self.fused_bn:

File [/usr/local/lib/python3.10/dist-packages/mmcv/ops/sparse_ops.py:99](https://vscode-remote+dev-002dcontainer-002b7b22686f737450617468223a222f686f6d652f7562756e74752f776f726b2f736d61727447617465222c2273657474696e6773223a7b22686f7374223a227373683a2f2f3139322e3136382e332e313737227d7d.vscode-resource.vscode-cdn.net/usr/local/lib/python3.10/dist-packages/mmcv/ops/sparse_ops.py:99), in get_indice_pairs(indices, batch_size, spatial_shape, ksize, stride, padding, dilation, out_padding, subm, transpose, grid)
     97     else:
     98         raise NotImplementedError
---> 99     return get_indice_pairs_func(indices, batch_size, out_shape,
    100                                  spatial_shape, ksize, stride, padding,
    101                                  dilation, out_padding, int(subm),
    102                                  int(transpose))
    103 else:
    104     if ndim == 2:

RuntimeError: [/tmp/mmcv/mmcv/ops/csrc/pytorch/cuda/sparse_indice.cu](https://vscode-remote+dev-002dcontainer-002b7b22686f737450617468223a222f686f6d652f7562756e74752f776f726b2f736d61727447617465222c2273657474696e6773223a7b22686f7374223a227373683a2f2f3139322e3136382e332e313737227d7d.vscode-resource.vscode-cdn.net/tmp/mmcv/mmcv/ops/csrc/pytorch/cuda/sparse_indice.cu) 126
cuda execution failed with error 2
DonghoonPark12 commented 1 year ago

@xiaomonv111 Same error. How could you solve it?

xiaomonv111 commented 1 year ago

no

17638759823 commented 11 months ago

This framework is too difficult to get started.