potato-disease-classification
potato-disease-classification copied to clipboard
Model Deployed to GCP gives wrong results
When it comes to Model Deployment To GCP. As user inputs, even meme images are giving output with 70% plus accuracy.
Could you please explain your trials?
I followed the tutorial and completed the model up to deployment with GCP. Model is running and producing accurate results according to given data. But still when testing the model for images outside the dataset it gives vaguely false interpretations of the disease. For instance suppose I upload an image of a meme, this model predicts a disease with 99% accuracy.