Predict is giving false positives

Part 1 2022
1

I am following lesson 1 “is it a bird” exercise. I am able to get it to train. When I gave a new bird image, it was able to predict. Then, I gave a forest image. Even that it predicted as true. Below is the code in the cell:
is_bird,_,probs = learn.predict(PILImage.create(‘/kaggle/working/640px-Quiver-Tree-Forest-2015.JPG’))
print(f"This is a: {is_bird}.“)
print(f"Probability it’s a bird: {probs[0]:.4f}”)

Here is the response

This is a: working.
Probability it’s a bird: 1.0000

Attached is the image.forestjpg

2

Hey Jay, :wave:

Here is a link of my bird or not bird notebook from the first lecture.

Do you want to try running that and see if you have any luck?

I’ve just double checked and it worked for me? :thinking:

3

Thanks for sharing. Will give it a try.

Also, for me the following line raised an exception

dls.show_batch(max_n=6)

It complained there are no batches. My problem might have been due to that.

4

@Grey-Shirt It works! I could fix issues in my code too! Thanks!!