nntrainer
nntrainer copied to clipboard
Details in error messages
Error messages should contain enough details to point out the location of error:
- Layer/LayerNode/LayerContext -> must include layer name in the error message
- Optimizer/OptimizerContext -> must include optimizer and weight name in the error message
- Tensor/VarGrad/Weight -> must include tensor name in the error message
- Classes -> must include class name and the high level task at which the error message cam
- *Later we can consider providing a call stack in the error message with debug mode
Apart from the correct identifier, possible hints/solution must be also added to the error message:
- give possible reasons why this error has occured (from the user side or from the internal behavior)
- give solutions for solving the error
:octocat: cibot: Thank you for posting issue #1803. The person in charge will reply soon.
additional list
- fc check if unit is empty at finalize, and issue proper warning