These are chat archives for beniz/deepdetect

27th
Apr 2016
Erkhemee
@erkhemee
Apr 27 2016 03:31
Hi again!
I'm trying to GoogLeNet cars classification service for test. Model file is here http://mmlab.ie.cuhk.edu.hk/datasets/comp_cars/googlenet_finetune_web_car_iter_10000.caffemodel
But, these files are old version of caffe. I upgraded those files with upgrade_net_proto tools
Erkhemee
@erkhemee
Apr 27 2016 03:36
And, My service creation command is here curl -X PUT "http://localhost:8080/services/car" -d "{\"mllib\":\"caffe\",\"description\":\"Cars classification\",\"type\":\"supervised\",\"parameters\":{\"input\":{\"connector\":\"image\",\"height\":224,\"width\":224},\"mllib\":{\"nclasses\":431}},\"model\":{\"repository\":\"/home/dd/models/cars\"}}"
Prediction cmd is here
curl -X POST "http://localhost:8080/predict" -d "{\"service\":\"car\",\"parameters\":{\"input\":{\"width\":224,\"height\":224},\"output\":{\"best\":3}},\"data\":[\"http://s1.cdn.autoevolution.com/images/gallery/MERCEDES-BENZ-S-class--W222--4853_7.jpg\"]}"
After that steps, dd daemon crashs
Erkhemee
@erkhemee
Apr 27 2016 03:41
am I missing something? Wrong parameters in service creation?
Emmanuel Benazera
@beniz
Apr 27 2016 06:43
@erkhemee hi, your deploy file is wrong and requires an in put MemoryData layer, but in all cases the server shouldn't crash. I would appreciate if you could open an issue. Please add the full server output as well. We'll deal with your input format file within the issue.