Encountering errors while training my own data using pointpillar documentation from matlab

Question

gaurav 2024-2-20

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/2084208-encountering-errors-while-training-my-own-data-using-pointpillar-documentation-from-matlab

回答： Yash Sharma 2024-7-22

doTraining= true 
if doTraining    
    [detector,info] = trainPointPillarsObjectDetector(cds,detector,options);
else
    pretrainedDetector = load('pretrainedPointPillarsDetector.mat','detector');
    detector = pretrainedDetector.detector;
end

while running this code from pointpillar documentation on my own data i am encountering this error

Error using dlnetwork/initialize
Invalid network.
Error in trainPointPillarsObjectDetector (line 210)
dlnet = initialize(dlnet);
Caused by:
Layer 'cnn|concatenate': Input size mismatch. Size of input to this layer is different from the expected input size.
Inputs to this layer:
from layer 'cnn|up1|bn' (size 69(S) × 125(S) × 128(C) × 1(B))
from layer 'cnn|up2|bn' (size 70(S) × 126(S) × 128(C) × 1(B))
from layer 'cnn|up3|bn' (size 72(S) × 128(S) × 128(C) × 1(B))

so how to overcome this and how to modify the layer and my data

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Yash Sharma 2024-7-22

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/2084208-encountering-errors-while-training-my-own-data-using-pointpillar-documentation-from-matlab#answer_1488816

Hi Gaurav,

The error you're encountering indicates a size mismatch in the inputs to the 'concatenate' layer in your PointPillars network. Here are concise steps to troubleshoot and resolve this issue:

1. Verify Input Data Dimensions

Ensure that the input point cloud dimensions match the expected input size for the PointPillars network.

2. Check Network Configuration

Review the configuration of the layers involved in the concatenation operation to ensure they produce compatible output sizes.

3. Data Preprocessing

Double-check your data preprocessing steps, such as voxelization and feature extraction, to ensure they are correctly implemented.