fix recording bug in after_train function #600
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
When I evaluate model in the
after_train
function and record results, the result will not be recorded inTensorboard
andmetrics.json
This is because detectron2.utils.events.
TensorboardXWriter
andJSONWriter
will check thetrainer.iter
, and each iter will only record once.If you do not set
trainer.iter+1
before executingafter_train
, the execution result will not be recorded, because the current trainer.iter has been used in the last stepSo I fixed this issue in
train_loop
function