Evaluation files test_evaluation_live.py and test_evaluation_csv.py that both test queries and log hyperparameters.
These files now use assert_test() instead of evaluate() because of the more digestible output and ability to log hyperparameters for model and prompt comparison. For this same reason, we had to split the live and csv evaluations into two separate files.
Evaluation files test_evaluation_live.py and test_evaluation_csv.py that both test queries and log hyperparameters. These files now use assert_test() instead of evaluate() because of the more digestible output and ability to log hyperparameters for model and prompt comparison. For this same reason, we had to split the live and csv evaluations into two separate files.
There is also an up to date README file.