-
Notifications
You must be signed in to change notification settings - Fork 105
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Rework module structure and packages #469
Conversation
It looks like some tests are not updated, at least these two:
|
Not related to the changes, but in this version, I cannot build the project using the command line |
Can you please elaborate on what should go to the |
…ies between "dataset" and "api" modules
Module "api" does not have any module-specific code, and "dataset" is used only for training now.
b141b25
to
8979363
Compare
No, that was my mistake. I fixed it and it should work now, thank you for pointing it out. |
Implementations of interfaces from the "api" package, platform-specific code, additional utilities such as Initially, I wanted to extract "preprocessing" module, but I noticed that a lot of classes also do not fit into the "api" category and also should be moved into a separate module. |
Current module structure has some problems:
TensorShape
orOperation
, would fit better in the "api" module and not in "dataset" module.To fix these problems, this PR introduces new module structure:
ImageRecognitionModelBase
, preprocessing implementations (including platform-dependent), utility classes, etc. It depends on "api" module.This PR also changes some packages, including moving "onnx" module classes to the package
org.jetbrains.kotlinx.dl.onnx
, andModelHub
and related classes fromkeras.loaders
to justloaders
(maybe here other package name likemodelhub
orhub
could be better?). Packages in "tensorflow" module are unchanged, but we may need to refactor them in the future releases.