rayguan97 / m3detr Goto Github PK

Hey Guys,

• I am trying to train the model using the given config file link_config_file just for 1 class.
• but I am getting this key error in the voxel_set_abstraction.py file.
• error message:
• File "..../tools/../pcdet/models/__init__.py", line 30, in model_func ret_dict, tb_dict, disp_dict = model(batch_dict) File "....../.pyenv/versions/3.7.7/envs/rp/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1051, in _call_impl return forward_call(*input, **kwargs) File "............../tools/../pcdet/models/detectors/ours_3d.py", line 11, in forward batch_dict = cur_module(batch_dict) File "............../.pyenv/versions/3.7.7/envs/rp/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1051, in _call_impl return forward_call(*input, **kwargs) File "............../tools/../pcdet/models/backbones_3d/pfe/voxel_set_abstraction.py", line 593, in forward batch_dict = self.reduce_points(batch_dict) File "............/tools/../pcdet/models/backbones_3d/pfe/voxel_set_abstraction.py", line 542, in reduce_points for bs_idx, roi in enumerate(batch_dict['batch_cls_preds']): KeyError: 'batch_cls_preds'
• I tried to print the batch_dict before the function call for reduce_points, and there is no key named as batch_cls_preds
• the dict I got is this BATCH_DICT: dict_keys(['points', 'frame_id', 'sample_idx', 'gt_boxes', 'use_lead_xyz', 'voxels', 'voxel_coords', 'voxel_num_points', 'metadata', 'batch_size', 'voxel_features', 'encoded_spconv_tensor', 'encoded_spconv_tensor_stride', 'multi_scale_3d_features', 'multi_scale_3d_strides', 'spatial_features', 'spatial_features_stride'])
• what could be the cause for this error? how can I solve this error?

Thanks in Advance!!

Will there be code availbe with soon?

Would be glad to see that :-)

Take care Jürgen

When training the M3DETR, there is an error which is : No key word batch_cls_preds happened in VoxelSetAbstractionTransFusionv5.

The specific code is as follows:

def reduce_points(self, batch_dict):

    batch_indices = batch_dict['points'][:, 0].long()

    masks = []
    for bs_idx, roi in enumerate(batch_dict['batch_cls_preds']):
        bs_mask = (batch_indices == bs_idx)
        pts = batch_dict['points'][bs_mask].unsqueeze(dim=1)[:, :, 1: 4]  # (N, 1, 3)
        s, _ = torch.max(batch_dict['batch_cls_preds'][bs_idx], dim=1)
        top, idx = torch.topk(s, self.topks)
        c = batch_dict['batch_box_preds'][bs_idx][idx][:, :3].unsqueeze(dim=0)
        dist = (pts - c)**2       

        dist, _ = dist.sum(dim=-1).min(dim=1)
        mask = (dist <= self.reduce_radius)
        masks.extend(mask)

    batch_dict['points'] = batch_dict['points'][masks]
    return batch_dict

It seems to me that the training phase should not be present: "batch_cls_preds," which should only be present in the inference phase. I would like to ask your opinion.

Do you have a training model for Kitti dataset

hello, I am very interested in your work. So I do exp. on KITTI dataset, but I found the result is lower than PV-RCNN and don't reproduce the same / close results as your paper both on R11 and R40 metrics. Could you please share the checkpoint? Thank you for your reply.

When I install SharedArray, it give me error below:

Building wheels for collected packages: SharedArray
Building wheel for SharedArray (pyproject.toml) ... error
ERROR: Command errored out with exit status 1:
command: 'F:\OutSourcing\3d_object_detection\M3DETR\venv\Scripts\python.exe' 'F:\OutSourcing\3d_object_detection\M3DETR\venv\lib\site-packages\pip_vendor\pep517\in_process_in_process.py' build_wheel 'C:\Users\CAIFEN~1\AppData\Local\Temp\tmpnoay6_36'
cwd: C:\Users\Cai Feng Lin\AppData\Local\Temp\pip-install-gc0uej1o\sharedarray_a6aff881d25c4fab845553e526ed24db
Complete output (12 lines):
running bdist_wheel
running build
running build_ext
building 'SharedArray' extension
creating build
creating build\temp.win-amd64-cpython-38
creating build\temp.win-amd64-cpython-38\Release
creating build\temp.win-amd64-cpython-38\Release\src
D:\Tools\VisualStudio\2019\Professional\VC\Tools\MSVC\14.29.30133\bin\HostX86\x64\cl.exe /c /nologo /O2 /W3 /GL /DNDEBUG /MD "-I
C:\Users\Cai Feng Lin\AppData\Local\Temp\pip-build-env-cqhwb4q3\overlay\Lib\site-packages\numpy\core\include" -IF:\OutSourcing\3d_
object_detection\M3DETR\venv\include -ID:\Tools\Python\Python38\include -ID:\Tools\Python\Python38\Include -ID:\Tools\VisualStudio
\2019\Professional\VC\Tools\MSVC\14.29.30133\ATLMFC\include -ID:\Tools\VisualStudio\2019\Professional\VC\Tools\MSVC\14.29.30133\in
clude "-IC:\Program Files (x86)\Windows Kits\NETFXSDK\4.8\include\um" "-ID:\Windows Kits\10\include\10.0.19041.0\ucrt" "-ID:\Windo
ws Kits\10\include\10.0.19041.0\shared" "-ID:\Windows Kits\10\include\10.0.19041.0\um" "-ID:\Windows Kits\10\include\10.0.19041.0
winrt" "-ID:\Windows Kits\10\include\10.0.19041.0\cppwinrt" /Tc.\src\map_owner.c /Fobuild\temp.win-amd64-cpython-38\Release.\src\map_owner.obj
map_owner.c
.\src\map_owner.c(19): fatal error C1083: Cannot open include file: 'sys/mman.h': No such file or directory
error: command 'D:\Tools\VisualStudio\2019\Professional\VC\Tools\MSVC\14.29.30133\bin\HostX86\x64\cl.exe' failed with exit code 2
ERROR: Failed building wheel for SharedArray
Failed to build SharedArray
ERROR: Could not build wheels for SharedArray, which is required to install pyproject.toml-based projects

So, how can I solve it?

Hi! Can you help with a few questions on positional embeddings:

1. Did you apply any positional embedding in the Transformer attention?
2. If so, how did you design the positional embeddings for different types of representation and multi-scale features, and how did you apply them (i.e. how the formula should write)?
3. If not, what's the consideration behind? Why is it not needed?

Thank you! I look forward to your reply and the code release.