Skip to content

Latest commit

 

History

History
48 lines (34 loc) · 2.37 KB

todo_list.md

File metadata and controls

48 lines (34 loc) · 2.37 KB
Raw

TODO

  • beautify UI
  • analyze and show model info on the index page, such as graph node number, model size etc.
  • ensure the model is fully loaded before modify() is called.
    • otherwise NameError: name 'onnx_modifier' is not defined error will be invoked.
  • support desktop application.
    • Windows
      • fix the shutdown issue
    • Linux
  • support more flexible downloading schema
    • As this request notes, the current downloading schema prevents onnx-modifier from being deployed remotely as a service.
  • support combine models (#63).
  • support user-defined input/output number when the type of node's input/output is list.
  • refine the code.
    • seperate the model graph rendering and editting into two classes in the js code for "cleaner" code.
    • make the code more readable, cleaner and consistent in the code format.
    • slim the code.
      • because some .js files (like electron.js and even python.js) in the static folder and electron.html in templates folder are legacies of Netron and can be further slimmed.
  • support adding model input/output node.
  • fix issue that "extra model inputs" emerges after deleting nodes. issue#12
  • support adding more complicated nodes (which has some simple parameters like reshape).
  • add shape inference feature (mentioned in this issue)
  • set cleanup and shape inference as user-switch-on-off function, rather than automatic post-process.

Some known reference issues/feature requests