Modelpublic

lcw99/polyglot-ko-5.8b-instruct-native-finetune

sourceHugging Faceapache-2.0updated 3y agoView on Hugging Face

0likes15downloads

Model Card

polyglot-5.8B Korean finetuned for instruction following

kobestcopa 0 shot result.(original polyglot-5.8b's score is 0.7745) | Task |Version| Metric |Value | |Stderr| |-----------|------:|--------|-----:|---|-----:| |kobestcopa| 0|acc |0.7970|± |0.0127| | | |macro_f1|0.7966|± |0.0127|