Features: Create an inference python file #35
Labels
Discussion
Extra attention is needed
Features
New feature or request
Inference
About benchmark inference file
Milestone
What
While we discussed about how we get model weights from directory when doing model soups, we noticed that we should create an inference(benchmark) python file.
Why
There are many benchmark githubs have an inference python file(or benchmark python file). Also, some of the functions we'll support need inference file.
How
Can reference validation function that is already existed in train.py.
Also, we'll reference lots of benchmark githubs from NeurIPS, MICCAI, etc.
Any opinions welcome!
The text was updated successfully, but these errors were encountered: