When I tried to train the model (after loading the 4 classes of training images - hdmi_male, thunderbolt_male, usb_male and vga_male), I get the following error.
"Training failed
Your model has failed training. Reason is: Cannot execute learning task. : Could not train classifier. Verify there are at least 10 positive training images for each class, and at least 10 other unique training images (inluding optional negative_examples). There is a minimum of 1 positive class. Not enough samples for training, class: hdmi_male has only 0 samples Click here for a summary."
This seems bizzare as
There are definitely more than 10 unique training and positive training images per class.
Definitely at least 1 positive class
Definitely more than 0 samples in the hdmi_male class.
Has anyone else encountered this issue?
Thanks,
Kartik
there are two steps for the data: first you upload the files, then you add them to the project, using what I call the ārefined hamburger menuā (3 dots instead of 3 bars): did you do the second step?
Yup! I did the steps you had outlined in the tutorial. Uploaded the files and then added them to the projectā¦after which I clicked on āTrain Modelā. See screenshot below showing models added to project.
hi! Iāve been trying to get into Watson Studio, to check the steps, but it just keeps spinning
The basic idea is to zip a folder of images that are NOT any of the positive classes: in the case of cables, you need images of things that arenāt cables, but look similar.
Upload it, then use the Negative (recommended) ābuttonā on the Default Custom Model.
