Closed yiyaxiaozhi closed 2 years ago
If training completed, you can replace rerank.traineval
with rerank.evaluate
to run only the evaluation. This will only work if all parts of training finished, which includes predictions on the dev set, for example.
I obtain the model checkpoint on Robust04 fold-2 by the following command:
python -m capreolus.run rerank.traineval with file=docs/reproduction/config_paradept.txt fold=s2
but while running the prediction, the progress is killed by accident. Therefore it did not display the test result. What command should I use to load the checkpoint and run the fold-2 test?