Resulting R_precision accuracy is always higher than the ground truth

Open RunqiWang77 opened this issue 1 year ago • 2 comments

Hello, I've tried many methods, but the resulting R_precision accuracy is always higher than the ground truth (GT). Is this situation normal, and is there a solution? Thank you very much.

Jun 21 '24 09:06 RunqiWang77

Try training from scratch.

Jul 06 '24 03:07 wangwenlonggg

Try training from scratch.

Hello,

Are you suggesting that we should retrain the Intergen model itself, or are you referring to retraining the feature extractor used for feature extraction?

Sep 16 '24 08:09 RunqiWang77