Watson Visual Recognition "Cannot execute learning task. : no classifier name given"

Question

Getting cURL error: {"code":400,"error":"Cannot execute learning task. : no classifier name given"}

Getting the same result whether I use the beta GUI tool or a cURL entry:

curl -X POST \
-F "Airplanes_positive_examples=@Airplanes.zip" \
-F "Biking_positive_examples=@Biking.zip" \
-F "GolfPuttingGreens_positive_examples=@GolfPuttingGreens.zip" \
-F "name=AllJpegClassifier" \
"https://gateway-a.watsonplatform.net/visual-recognition/api/v3/classifiers?api_key={my-api-key}&version=2016-05-20"

I have read all previous SO questions for this problem and made sure of the following:

• Classifier name is alphanumeric only
• Zip filenames are alphanumeric only
• Image filenames are alphanumeric with _ - . only
• Zip files contain between 27 and 49 images each
• All image files are the same format (JPEG)
• All images conform to pixel size and file size limits

Answer 1

Your command looks fine, and when I try it with my API key and my own zip files, it works. So I suspect there is something in your zip files that the system is having trouble with. If you could provide the "owner" guid field (also called you instance-id) I could look into our logs to try to diagnose it. This is displayed when you do a GET /classifiers/{cid} of an existing classifier. Alternatively, you could let me know one of your other existing classifier_ids

Another way would be if you could open a Bluemix support ticket and include copies of the zip files which you're using in this example. Then we can reproduce the problem.