These are chat archives for beniz/deepdetect

6th
Jan 2018
Yili Zhao
@panovr
Jan 06 2018 02:01
@beniz When I finetuning resnet50, there was an error:

E0106 09:58:23.671247 2680 caffelib.cc:1016] Error while proceeding with test forward pass
INFO - 09:58:23 - Opened lmdb /home/ylzhao/models/yub_resnet50/test.lmdb
ERROR - 09:58:23 - service imgserv training status call failed

ERROR - 09:58:23 - {"code":500,"msg":"InternalError","dd_code":1007,"dd_msg":"src/caffe/util/im2col.cu:61 / Check failed (custom): (error) == (cudaSuccess)"}

I use CUDA 9.1, is this a bug in deepdetect's caffe?
Emmanuel Benazera
@beniz
Jan 06 2018 07:09
DD works fine with cuda9.1, I believe you should look at the FAQ, https://deepdetect.com/overview/faq/
Yili Zhao
@panovr
Jan 06 2018 10:07
@beniz you mean that error was caused by insufficient GPU memory, and I should decrease batch_size?