Is there a way for me to slightly increase the size of the dense_vector datatype introduced in the Elasticsearch 7.0.0 beta?
For my application I need to store vectors with 512 dimensions, but the documentation states "The maximum number of dimensions that can be in a vector should not exceed 500".
When I attempt to store a vector of 512 dimensions I get a "RequestError(400, mapper_parsing_exception, failed to parse)" which seems to be due to this limitation.
A different approach would be that each object has 512 fields containing a single float, each representing one dimension. This seems like a less elegant solution. I also can't use an array of floats as the order of the dimensions is important.