You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi! I am training the MVXNet, and got a question if there is the best model save somewhere. I mean it saves the model from the current epoch, but I don't see how to get the best model, because its not always the last one?
Can you please help! Thank you a lot in advance.
The text was updated successfully, but these errors were encountered:
Currently we only save the link to the latest model. You can refer to gather_models.py for the way to find the best model according to a specific metric, which is usually used to find the logs/checkpoints of the best model when we set a new benchmark.
* check in python demos
* check in text detector python demo
* check in roatated object python demo
* check in pose python demo
* ignore the output class number when testing metrics with sdk as a backend
* fix object_detection
* rollback segmentation_model and python/segmentor.cpp
Hi! I am training the MVXNet, and got a question if there is the best model save somewhere. I mean it saves the model from the current epoch, but I don't see how to get the best model, because its not always the last one?
Can you please help! Thank you a lot in advance.
The text was updated successfully, but these errors were encountered: