testing on own dataset

[Puneesh6280]

I tried adding an own tif file to test folder, but showing me an error.
Is there a way to add and test on our own dataset?

[souvikmajumder26]

Did you try to test the model with the sample dataset mentioned in the project on your own tif file? In that case the error might be due to the image bands/channels mismatch. If you build a model solely on your own dataset while tweaking the code (if required) as per the tif file bands/channels it should be able to run with a similar test file.

[AdityaPrakash-Evigway]

how to inference the code without providing the masks. I wanted to test the model on my dataset which is not labeled.

Did you try to test the model with the sample dataset mentioned in the project on your own tif file? In that case the error might be due to the image bands/channels mismatch. If you build a model solely on your own dataset while tweaking the code (if required) as per the tif file bands/channels it should be able to run with a similar test file.

[andrewp-CWF]

You can comment out lines 98-106 and line 194 in src/test.py to predict on your own images without a ground truth mask.

how to inference the code without providing the masks. I wanted to test the model on my dataset which is not labeled.

Did you try to test the model with the sample dataset mentioned in the project on your own tif file? In that case the error might be due to the image bands/channels mismatch. If you build a model solely on your own dataset while tweaking the code (if required) as per the tif file bands/channels it should be able to run with a similar test file.