Originally posted by **jsvan** July 3, 2024
Hello,
Could you please share the testing scripts for the Narrative/QuALITY/Qasper scores? I'd like to run some experiments on how quantization/reducing embedding dimensions affect performance. However running these look non trivial because it's information retrieval/generation etc., and I'm trying to fit this in while procrasting from other work I should be doing.
Thank you!
Discussed in https://github.com/parthsarthi03/raptor/discussions/47