Open seungyun1 opened 2 years ago
Seems to be a compilation error for scatter_points_cuda
. Could you try the solutions here https://github.com/open-mmlab/mmdetection3d/issues/362?
Seems to be a compilation error for
scatter_points_cuda
. Could you try the solutions here open-mmlab#362?
Thanks to you, I was able to solve it.
Thank you very much.
Seems to be a compilation error for
scatter_points_cuda
. Could you try the solutions here open-mmlab#362?
Finally, there is a question.
where can I get 'fusion_voxel0075.pth' ?
there is a error
OSError: checkpoints/fusion_voxel0075_R50.pth is not a checkpoint file
When I run python tools/train.py configs/transfusion_nusc_voxel_LC.py
Seems to be a compilation error for
scatter_points_cuda
. Could you try the solutions here open-mmlab#362?Finally, there is a question.
where can I get 'fusion_voxel0075.pth' ?
there is a error
OSError: checkpoints/fusion_voxel0075_R50.pth is not a checkpoint file
When I run
python tools/train.py configs/transfusion_nusc_voxel_LC.py
Just set in config file's 'load_from = None' ?
Glad you solve the compile issue. For this question, I am afraid I can not share the model checkpoints due to the policy of Huawei. So you need to train both the TransFusion-L and TransFusion by yourself. Basically:
fusion_voxel0075.pth
as the load_from
key for TransFusion.You can find the detail in https://github.com/XuyangBai/TransFusion/blob/master/configs/nuscenes.md
Glad you solve the compile issue. For this question, I am afraid I can not share the model checkpoints due to the policy of Huawei. So you need to train both the TransFusion-L and TransFusion by yourself. Basically:
- Train transfusion_nusc_voxel_L.py
- Choose a 2D backbone. For nuscenes dataset, you can directly use the model provided by mmdet3d. Then you can combine the pretrained TransFusionL and 2D backbone to get the
fusion_voxel0075.pth
as theload_from
key for TransFusion.- Train transfusion_nusc_voxel_LC.py
You can find the detail in https://github.com/XuyangBai/TransFusion/blob/master/configs/nuscenes.md
Thank you very much for your kind reply!!
Seems to be a compilation error for
scatter_points_cuda
. Could you try the solutions here open-mmlab#362?Thanks to you, I was able to solve it.
Thank you very much.
hello, I have the same error. How did you solve it? thanks a lot !
Seems to be a compilation error for
scatter_points_cuda
. Could you try the solutions here open-mmlab#362?Thanks to you, I was able to solve it. Thank you very much.
hello, I have the same error. How did you solve it? thanks a lot ! 找到文件--找到需要修改的行--把coors_id_argsort加上{}
/TransFusion/mmdet3d/ops/voxel/src/scatter_points_cuda.cu
//coors_map.indexput(coors_id_argsort, coors_map_sorted); coors_map.indexput({coors_id_argsort}, coors_map_sorted);
Seems to be a compilation error for
scatter_points_cuda
. Could you try the solutions here open-mmlab#362?Thanks to you, I was able to solve it. Thank you very much.
hello, I have the same error. How did you solve it? thanks a lot !
Hello, I also encountered the same problem. How do I get the checkpoints file fusion_model.pth after I have completed the blood transfusion-L training? Please guide me.Thank you!
Thanks for your error report and we appreciate it a lot.
Checklist
Describe the bug A clear and concise description of what the bug is.
Can you give me some advice?
The error is still occurring for me.
I tried to run 'pip install -v -e.', but there is an error like below.
I don't know what the problem is. I'd appreciate it if anyone could tell me how to solve it.
I want train for using transfusion model but I don't know how to solve this problem
two gpu gpu 0 : NVIDIA TITAN X gpu 1 : NVIDIA TITAN RTX Reproduction
Environment
python mmdet3d/utils/collect_env.py
to collect necessary environment infomation and paste it here.TorchVision: 0.11.3 OpenCV: 4.5.5 MMCV: 1.5.0 MMCV Compiler: GCC 7.5 MMCV CUDA Compiler: 11.1 MMDetection: 2.24.1 MMSegmentation: 0.24.1 MMDetection3D: 1.0.0rc2+76e351a spconv2.0: False
A placeholder for trackback.