Error "Request Entity Too Large" when trying to train Visual Recognition model

0
votes

I'm receiving this error when trying to train my first custom model:

Unauthorized: Access is denied due to invalid credentials. Ensure that the Watson Visual Recognition service has not been deleted, and that the quota limit for your plan has not been reached.

I haven't done much at this point. I also am not accessing Visual Recognition through anything else, I'm just trying to get this model trained. I'm not sure what credentials or quota limit would have to do with this. Anyone else have any experience with this issue?

Back story: I had originally uploaded 73 zip files (all >10 files inside, each ~5MB, 1342 images total), but I was catching errors when trying to train, including a "request too large". So I declassified 70 of them, and now I'm just trying to train this model with 3 categories. Now I'm getting the "Unauthorized" message I originally mentioned. I had given some time in between pressing the train button (hours) to maybe prevent any backlog of requests from my part.

1
ibm-watsoncoremlvisual-recognitionwatson-studio
Is this the same service instance that you used before or a new one? Also check the plan that the instance is using, is it now on lite. Check the endpoint url, depending on when you created the old instance, this might have changed. – chughts

1 Answers

0
votes

Try a command like this from our docs to use the general classifier, to test whether your apikey is valid, or if you have perhaps used up the free (Lite) plan allocation for the current month:

curl -u "apikey:{apikey}" "https://gateway.watsonplatform.net/visual-recognition/api/v3/classify?url=https://watson-developer-cloud.github.io/doc-tutorial-downloads/visual-recognition/fruitbowl.jpg&version=2018-03-19"

For other languages, check the API reference here: https://console.bluemix.net/apidocs/visual-recognition#classify-an-image