I am trying to create a custom image classifier on Jupyter Notebooks via Python 2.7 using the VisualRecognitionV3 class from the python Watson developer cloud service. I created a Watson class to implement the create_classifier(), get_classifier(), delete_classifier(), and classify() functions. I had an api key from before but now after running the previous functions I have the following errors:
running get_classifier(classifier_id = ''):
Output: {u'classifiers': []}
running create_classifier():
Output:
WatsonException: Error: Cannot execute learning task. : this plan instance can have only 1 custom classifier(s), and 1 already exist., Code: 400
running delete_classifier(classifier_id = ''):
Output:
WatsonException: Unknown error
running classify(classifier_id = '')
Output:
{
"images": [
{
"image": filepath,
"classifiers": [
{
"classes": [
{
"score": 0.738,
"class": "open-end wrench",
"type_hierarchy": "/tool/open-end wrench"
},
{
"score": 0.785,
"class": "wrench"
},
...
{
"score": 0.644,
"class": "dark red color"
}
],
"classifier_id": "default",
"name": "default"
}
]
}
],
"custom_classes": 0,
"images_processed": 1
}
I was supposed to get two classes for the bottom but I have zero classes. I have tried using another API key via a 30-day trial but I am getting an Error 403: invalid API key. Does anyone know how to eliminate the current class without a classifier_id and generate a new one, or use the old one to generate a working one? Any ideas would be much appreciated. Thanks!