VisualRecognitionV3 new classifier status = fehlgeschlagen

Question

Ich habe den folgenden Python SDK-Beispielcode und Eingabedateien in Bluemix, um neue Klassifizierer zu erstellen. Hier ist der Code:

visual_recognition = VisualRecognitionV3('2016-05-20',api_key=apikey)with 
open('./cars.zip', 'rb') as cars, open('./trucks.zip', 'rb') as trucks: 
      print(json.dumps(visual_recognition.create_classifier('Cars', 
      positive_examples=cars, negative_example=trucks), indent=2)) 
      print(json.dumps(visual_recognition.list_classifiers(), indent=2)) 

RUNLOG zeigt Status = fehlgeschlagen.

Was fehlt mir?

17/6/2016 19:50:03 OUT App { 
17/6/2016 19:50:03 OUT App "owner": "885160bd-637d-4f8f-b268-fed5df81d672", 
17/6/2016 19:50:03 OUT App "classifier_id": "Cars_2004369051", 
17/6/2016 19:50:03 OUT App "created": "2016-06-17T19:50:03.301Z", 
17/6/2016 19:50:03 OUT App "name": "Cars", 
17/6/2016 19:50:03 OUT App "classes": [], 
17/6/2016 19:50:03 OUT App "status": "training" 
17/6/2016 19:50:03 OUT App } 
17/6/2016 19:50:05 OUT App { 
17/6/2016 19:50:05 OUT App "classifiers": [ 
17/6/2016 19:50:05 OUT App { 
17/6/2016 19:50:05 OUT App "name": "Cars", 
17/6/2016 19:50:05 OUT App "classifier_id": "Cars_933017903", 
17/6/2016 19:50:05 OUT App "status": "failed" 
17/6/2016 19:50:05 OUT App }, 

Answer 1

1. Dieser arbeitete für mich:

   import json 
   from watson_developer_cloud import VisualRecognitionV3 as vr 
   instance =vr(api_key='apikey', version='2016-05-20') 
   with open('./trucks.zip', 'rb') as trucks, \ 
       open('./cars.zip', 'rb') as cars: 
       visual_recognition.create_classifier('CarsvsTrucks', cars_positive_examples=cars, negative_examples=trucks)