When I trained my deep model for some iterations, I got some warning about the data cache because of memory limitation. See the sentence below for the warning.
W0321 10:26:16.957023 93510 data_reader.cpp:197] Data Reader cached 1148000 records so far but it can't continue because it used more than half of swap buffer. Free swap memory left: 36476575744 of total 73014435840. Cache and shuffling are now disabled.
Will the cache and shuffling restart after some further iterations during the training?
Hi anyone,
When I trained my deep model for some iterations, I got some warning about the data cache because of memory limitation. See the sentence below for the warning.
W0321 10:26:16.957023 93510 data_reader.cpp:197] Data Reader cached 1148000 records so far but it can't continue because it used more than half of swap buffer. Free swap memory left: 36476575744 of total 73014435840. Cache and shuffling are now disabled.
Will the cache and shuffling restart after some further iterations during the training?
Thanks for your help.
caffe version: caffe-0.16.6 data type: LMDB