Check failed: MaxBottomBlobs() >= bottom.size() (2 vs. 3) L1Loss Layer takes at most 2 bottom blob(s) as input.

[TrackingBird]

Dear sir, have you ever meet the error when you training, re-train or fine-tuning your code with the kitti-dataset. Here follows the error. Hope for your help. Thanks very much.

I0507 15:28:09.845563 30538 net.cpp:91] Creating Layer flow_loss1
I0507 15:28:09.845571 30538 net.cpp:426] flow_loss1 <- blob66_NegReLU6_0_split_0
I0507 15:28:09.845584 30538 net.cpp:426] flow_loss1 <- blob65
I0507 15:28:09.845593 30538 net.cpp:426] flow_loss1 <- blob65Confidence
I0507 15:28:09.845618 30538 net.cpp:400] flow_loss1 -> flow_loss1
F0507 15:28:09.845647 30538 layer.hpp:404] Check failed: MaxBottomBlobs() >= bottom.size() (2 vs. 3) L1Loss Layer takes at most 2 bottom blob(s) as input.
*** Check failure stack trace: ***
    @     0x7f437a19e5cd  google::LogMessage::Fail()
    @     0x7f437a1a0433  google::LogMessage::SendToLog()
    @     0x7f437a19e15b  google::LogMessage::Flush()
    @     0x7f437a1a0e1e  google::LogMessageFatal::~LogMessageFatal()
    @     0x7f437aa53ff7  caffe::Layer<>::CheckBlobCounts()
    @     0x7f437aa53113  caffe::Layer<>::SetUp()
    @     0x7f437ab6fa08  caffe::Net<>::Init()
    @     0x7f437ab6dd51  caffe::Net<>::Net()
    @     0x7f437aba1101  caffe::Solver<>::InitTrainNet()
    @     0x7f437aba08ea  caffe::Solver<>::Init()
    @     0x7f437aba0334  caffe::Solver<>::Solver()
    @     0x7f437abae31b  caffe::SGDSolver<>::SGDSolver()
    @     0x7f437abb2e04  caffe::AdamSolver<>::AdamSolver()
    @     0x7f437abb3d38  caffe::Creator_AdamSolver<>()
    @           0x420d66  caffe::SolverRegistry<>::CreateSolver()
    @           0x41c0c4  train()
    @           0x41e639  main
    @     0x7f4378c52830  __libc_start_main
    @           0x41ae39  _start
    @              (nil)  (unknown)

[nikolausmayer]

You have three input blobs for an L1Loss layer. The layer does not like this and complains: L1Loss Layer takes at most 2 bottom blob(s) as input.

This has nothing to do with KITTI specifically.

[nikolausmayer]

(closed due to inactivity)