How to fix debuged CNN model in Keras
Data Science Asked by DispooL on December 23, 2020
I’ve trained a CNN model based on landscape and pool images. The main purpose is to make model classify if the image contains a pool or not. The accuracy on the test set was about 94% but when I used my own image to test, sometimes model did wrong predictions then I implemented GradCam and realized that the model looks only at the edges of the pool when I would classify the pool by the water. How can I fix my model so it could make better predictions and look not only at the edges?
One Answer
I solved the problem, I added an additional conv layer with 64 feature maps and increased size of dense layer from 512 to 1024 also I trained the model with the increased number of epochs(from 50 to 150). The result is impressive! Everything as I wanted.
Answered by DispooL on December 23, 2020
Add your own answers!
Ask a Question
Get help from others!
Recent Questions
- How can I transform graph image into a tikzpicture LaTeX code?
- How Do I Get The Ifruit App Off Of Gta 5 / Grand Theft Auto 5
- Iv’e designed a space elevator using a series of lasers. do you know anybody i could submit the designs too that could manufacture the concept and put it to use
- Need help finding a book. Female OP protagonist, magic
- Why is the WWF pending games (“Your turn”) area replaced w/ a column of “Bonus & Reward”gift boxes?
Recent Answers
- Lex on Does Google Analytics track 404 page responses as valid page views?
- Jon Church on Why fry rice before boiling?
- Peter Machado on Why fry rice before boiling?
- Joshua Engel on Why fry rice before boiling?
- haakon.io on Why fry rice before boiling?