4

after troubleshooting everything i thought my csv was perfect, any idea what this error means? Found nothing searching.

This is what my csv looks like: enter image description here but with 84k rows and 4 different labels.

The error:

Error: gs://retina-261312-vcm/csv/finalbookLOOL.csv line 105: The label is already a ground truth.
Error: gs://retina-261312-vcm/csv/finalbookLOOL.csv line 20051: The label is already a ground truth.
Error: gs://retina-261312-vcm/csv/finalbookLOOL.csv line 94: The label is already a ground truth.
Error: gs://retina-261312-vcm/csv/finalbookLOOL.csv line 99: The label is already a ground truth.
Error: gs://retina-261312-vcm/csv/finalbookLOOL.csv line 98: The label is already a ground truth.
Error: gs://retina-261312-vcm/csv/finalbookLOOL.csv line 92: The label is already a ground truth.

any help is appreciated

Faddy Sunna
  • 61
  • 3

1 Answers1

2

Ground truth means what you measured for your target variable for the training and testing examples. It means a target you want your model to predict.

Possibilities are some noise found, but it force to use model to predict. The ground truth will appear.

Please check it the ground truth explanation here

kertijayancom
  • 31
  • 2
  • 1
    I dont understand, I am just trying to create an image dataset on automl vision. I am setting the labels as in the documentation. Why am i getting this error? – Faddy Sunna Dec 13 '19 at 21:43
  • 2
    just remove the line that contain error. it appear because some of your testing data is perfectly simlar to training data in automl model. – kertijayancom Dec 13 '19 at 21:52
  • I remove some, and it finds other lines to trigger the error. I have not yet separated the data into test/train/val. I am just uploading it with labels. Is it the folder structure? I have, BUCKET/images/label1/images .. BUCKET/images/label2/images etc . – Faddy Sunna Dec 13 '19 at 22:56
  • Could You resolve this issue? I am facing the same issue as well – Rajarshi Bhadra Feb 02 '20 at 12:28