search

guigzzz / Keras-Yolo-v2

Keras re-implementation of Yolo v2 Object Detection
29 stars 8 forks source link

Where is fetch_bounding_boxes_and_labels defined? #1

Closed martinbel closed 5 years ago

guigzzz commented 5 years ago

It isn't defined anywhere, it is a placeholder. I explain the format of bounding_boxes and labels in the readme, so as long as your data follows that format, the code should work the same.

I also wanted to note that I haven't actually used this code to train a full model, so there is no guarantee that it will actually work (even though I was able to overfit on a simple example).

However inference should work fine, so you should be able train your model with darknet, and then infer using this code

martinbel commented 5 years ago

It's a pity the function to read at least one common annotation format isn't implemented. You can probably adapt this to fit the project without much effort. Beside this, everything looks great! The code is very intuitive and simple to extend.

guigzzz commented 5 years ago

I would guess that tf's standard library has helper functions to parse common annotation formats, if not it's not very difficult to find ressources online. This repo was a learning experience for me, and not much more. I'm happy you like the code. If you find any bugs/mistakes don't hesitate to create a PR :)