DaloroAT/first_breaks_picking

didn't understand how to pick first break?

Closed this issue · 5 comments

ad1995-geek commented

could you help me out , how to pick first break? I have read all of the codes. but only functions are defined. Didn't understand what are the functions are to be called ?

DaloroAT commented

Yes, of course.

First breaks are calculated during training to keep track of metrics. This is done using function here. This function does several calculations based on the last layer's featuremap.

To get a vector with the values of the first breaks, use this function.

poprafka commented

Hi DaloroAT, really interesting code and a nice readme! However, I'd like to add a vote to a better documentation and how to use/test the code. An example code introducing and demonstrating the expected input for the individual steps would be very appreciated.

  • 👍3
DaloroAT commented

Hi @poprafka . Nice to hear that someone is interested in improving the project. In fact, I did this project when I wanted to go into deep learning, it was part of my CV. The last commit was more than two years ago when I was hired))

If you need usage documentation, I'll add it as soon as free time becomes available.

Also I continue the idea of ​​this project in a private repository. It changed the architecture, the network was trained on real data, and a FastAPI web server was also written to serve the model. If you are interested, write to at.daloro@gmail.com.

  • 👍1
HHTC0 commented

Hi, thank you for the code. As you mentioned above, did you had a chance to write down document on how to use the code?

DaloroAT commented

Hello @ad1995-geek @poprafka @HHTC0 !
I've finally updated the repo, added tutorials and model weights, and released new features.

You can find updates on two URLs:
https://github.com/DaloroAT/first_breaks_picking
https://pypi.org/project/first-breaks-picking/

  • 👍1