These are chat archives for beniz/deepdetect

18th
Nov 2017
jubeenshah
@jubeenshah
Nov 18 2017 10:46
Hello, I was finally able to start the deep detect server, and start training an image classifier. Thanks a lot for the support. However, I'm stuck somewhere. I'm able to start a service, and start training on a small dataset, but I get the following exception. This network produces output loss1/loss1 INFO - 13:07:19 - This network produces output loss1/top-1 INFO - 13:07:19 - This network produces output loss1/top-5 INFO - 13:07:19 - This network produces output loss2/loss1 INFO - 13:07:19 - This network produces output loss2/top-1 INFO - 13:07:19 - This network produces output loss2/top-5 INFO - 13:07:19 - This network produces output loss3/loss3 INFO - 13:07:19 - This network produces output loss3/top-1 INFO - 13:07:19 - This network produces output probt INFO - 13:07:19 - Network initialization done. E1118 13:07:19.111397 15993 caffelib.cc:785] exception while forward/backward pass through the network
for which reason i'm not able to predict...
I run the following command for prediction

`ubuntu@ubuntu:~/Documents/Jubeen/training/imagenet_downloader$ curl -X POST "http://localhost:8080/predict" -d '{

   "service":"imageserv_test_4",
   "parameters":{
     "input":{
       "width":224,
       "height":224
     },
     "output":{
       "best":3
     }
   },
   "data":["/home/ubuntu/Documents/Jubeen/training/ambulance.jpg"]
 }'

{"status":{"code":500,"msg":"InternalError"}}ubuntu@ubuntu:~/Documents/Jubeen/training/imagenet_downloader$
`

and get the corresponding error, while deepdect server gives me this error...

`ERROR - 16:18:11 - service imageserv_test_4 mllib internal error: no model in /home/ubuntu/Documents/Jubeen/training/test_dir/data for initializing the net

ERROR - 16:18:11 - Sat Nov 18 16:18:11 2017 IST - 127.0.0.1 "POST /predict" 500 0
`

jubeenshah
@jubeenshah
Nov 18 2017 10:52
the command I run for setting up the service is
 curl -X PUT "http://localhost:8080/services/imageserv_test_4" -d '{
       "mllib":"caffe",
       "description":"image classification service",
       "type":"supervised",
       "parameters":{
         "input":{
           "connector":"image",
           "width":224,
           "height":224
         },
         "mllib":{
           "template":"googlenet",
           "nclasses":5
         }
       },
       "model":{
         "templates":"/home/ubuntu/Documents/Jubeen/training/test_dir/caffe",
         "repository":"/home/ubuntu/Documents/Jubeen/training/test_dir"
       }
     }'
and the command I run for starting to train is
curl -X POST "http://localhost:8080/train" -d '{
>        "service":"imageserv_test_2",
>        "async":true,
>        "parameters":{
>          "mllib":{
>            "gpu":true,
>            "net":{
>              "batch_size":32
>            },
>            "solver":{
>              "test_interval":500,
>              "iterations":30000,
>              "base_lr":0.001,
>              "stepsize":1000,
>              "gamma":0.9
>            }
>          },
>          "input":{
>            "connector":
>            "image",
>            "test_split":0.1,
>            "shuffle":true,
>            "width":224,
>            "height":224
>          },
>          "output":{
>            "measure":["acc","mcll","f1"]
>          }
>        },
>        "data":["/home/ubuntu/Documents/Jubeen/training/test_dir"]
>      }'
"service":"imageserv_test_4"