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
We should instead dynamically choose the default based on the data type. Float-based vectors can use L2. Binary vectors should use hamming.
Also, we should fix the error messsage: If given a wrong distance type, we should return Error::InvalidInput and the message should say that the distance function doesn't support the data type. Note that the current message doesn't say which distance function has been selected.
The text was updated successfully, but these errors were encountered:
The default distance type is L2, which works for float-based vectors. But for binary vectors, this results in hard-to-understand error:
We should instead dynamically choose the default based on the data type. Float-based vectors can use L2. Binary vectors should use hamming.
Also, we should fix the error messsage: If given a wrong distance type, we should return
Error::InvalidInput
and the message should say that the distance function doesn't support the data type. Note that the current message doesn't say which distance function has been selected.The text was updated successfully, but these errors were encountered: