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
According to the research paper, the following line : "We directly adopted the evaluator again for knowledge strips filtering, and the top-k is set to 5, filter threshold as -0.5", mentions that a threshold of -0.5 was used to filter the strips, but it is nowhere to be found in the internal_knowledge_preparation.py script.
The text was updated successfully, but these errors were encountered:
We chose the top-k filtering strategy in the repo. You can change the strategy by modifying the function select_relevants and setting the threshold. We will also update the second strategy later.
According to the research paper, the following line : "We directly adopted the evaluator again for knowledge strips filtering, and the top-k is set to 5, filter threshold as -0.5", mentions that a threshold of -0.5 was used to filter the strips, but it is nowhere to be found in the internal_knowledge_preparation.py script.
The text was updated successfully, but these errors were encountered: