This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TDA2EG-17: caffe-jacinto - train_imagenet_classification.sh cannot read lmdb dataset

Part Number: TDA2EG-17

Hi,

I've modified the train_imagenet_classification.sh file to train on my custom lmdb dataset in a docker container with gpu.

But when I execute the script, there is an error:

F0325 02:25:37.788183  4729 db_lmdb.hpp:16] Check failed: mdb_status == 0 (2 vs. 0) No such file or directory
*** Check failure stack trace: ***
    @     0x7f85e11f45cd  google::LogMessage::Fail()
    @     0x7f85e11f6433  google::LogMessage::SendToLog()
    @     0x7f85e11f415b  google::LogMessage::Flush()
    @     0x7f85e11f6e1e  google::LogMessageFatal::~LogMessageFatal()
    @     0x7f85e1de77e0  caffe::db::LMDB::Open()
    @     0x7f85e2282dc7  caffe::DataReader<>::InternalThreadEntryN()
    @     0x7f85e22753f0  caffe::InternalThread::entry()
    @     0x7f85e227745b  boost::detail::thread_data<>::run()
    @     0x7f85d0c7e5d5  (unknown)
    @     0x7f85b943c6ba  start_thread
    @     0x7f85df96541d  clone
    @              (nil)  (unknown)

I've searched online, and tried the following solutions but none worked:

Set CAFFE_ROOT to /caffe-jacinto

Add /caffe-jacinto/python to PYTHONPATH

use absolute path of the lmdb directory in ./models/image_classification.py instead of relative path

change the permission of the mdb file so it can be read, written and executed

check data.mdb and lock.mdb, both are in the directory and not empty

Thank you for your time and assistance.

Kind regards,

Wenbin