DIGITS: Error Code -11 Training DetectNet on KITTI

I get this error repeatedly while trying to train DetectNet on the KITTI dataset.

Iteration 632 (0.410104 iter/s, 192.634s/79 iter), loss = 20.8109
Train net output #0: loss_bbox = 7.84001 (* 2 = 15.68 loss)
Train net output #1: loss_coverage = 16.201 (* 1 = 16.201 loss)
Iteration 632, lr = 0.0001
Snapshotting to binary proto file snapshot_iter_638.caffemodel
Snapshotting solver state to binary proto file snapshot_iter_638.solverstate
Iteration 638, Testing net (#0)
Ignoring source layer train_data
Ignoring source layer train_label
Ignoring source layer train_transform

I am following this example and everything else in DIGITS seems to work just fine so I believe that it is related to opencv but am unsure of how to proceed:

https://github.com/NVIDIA/DIGITS/tree/master/examples/object-detection

About this issue

Original URL
State: closed
Created 8 years ago
Comments: 15 (5 by maintainers)

Most upvoted comments

The training fails at the same point but now the error changes to Error Code 1 instead of -11. This are the last few lines of the log. I’m not sure what this /home/travis directory is, I don’t have one.

I1207 14:25:54.376837  5216 solver.cpp:479] Snapshotting to binary proto file snapshot_iter_638.caffemodel
I1207 14:25:54.523226  5216 sgd_solver.cpp:273] Snapshotting solver state to binary proto file snapshot_iter_638.solverstate
I1207 14:25:54.571719  5216 solver.cpp:362] Iteration 638, Testing net (#0)
I1207 14:25:54.571735  5216 net.cpp:723] Ignoring source layer train_data
I1207 14:25:54.571738  5216 net.cpp:723] Ignoring source layer train_label
I1207 14:25:54.571740  5216 net.cpp:723] Ignoring source layer train_transform
OpenCV Error: The function/feature is not implemented (Unknown/unsupported array type) in type, file /home/travis/miniconda/conda-bld/work/opencv-3.1.0/modules/core/src/matrix.cpp, line 1879
OpenCV Error: The function/feature is not implemented (Unknown/unsupported array type) in type, file /home/travis/miniconda/conda-bld/work/opencv-3.1.0/modules/core/src/matrix.cpp, line 1879
Traceback (most recent call last):
File "/home/mike/caffe/python/caffe/layers/detectnet/clustering.py", line 133, in forward
bbox = cluster(self, data0, bottom[1].data)
File "/home/mike/caffe/python/caffe/layers/detectnet/clustering.py", line 227, in cluster
boxes_cur_image = vote_boxes(propose_boxes, propose_cvgs, mask, self)
File "/home/mike/caffe/python/caffe/layers/detectnet/clustering.py", line 193, in vote_boxes
self.gridbox_rect_eps)
cv2.error: /home/travis/miniconda/conda-bld/work/opencv-3.1.0/modules/core/src/matrix.cpp:1879: error: (-213) Unknown/unsupported array type in function type

MikepX99 on Dec 7, 2016

Reinstalling everything and removing Anaconda allowed the training to make it past epoch 1 and the validation test. Hopefully DetectNet will properly train now.

MikepX99 on Dec 7, 2016