logical-reasoning / results /mgtv-llama3_p1_r2_full_metrics.csv
dh-mc's picture
llama3 en
58a3992
raw
history blame
719 Bytes
epoch,model,accuracy,precision,recall,f1
0,shenzhi-wang/Llama3-8B-Chinese-Chat_torch.bfloat16,0.7836666666666666,0.7667122897184859,0.7929173693086004,0.7679400621793133
1,shenzhi-wang/Llama3-8B-Chinese-Chat/checkpoint-175_torch.bfloat16,0.7706666666666667,0.8072750943858197,0.7706666666666667,0.7835719791561528
2,shenzhi-wang/Llama3-8B-Chinese-Chat/checkpoint-350_torch.bfloat16,0.724,0.8118050163437011,0.724,0.7562266825513707
3,shenzhi-wang/Llama3-8B-Chinese-Chat/checkpoint-525_torch.bfloat16,0.6756666666666666,0.7811762160181578,0.6756666666666666,0.7108457483297581
4,shenzhi-wang/Llama3-8B-Chinese-Chat/checkpoint-700_torch.bfloat16,0.6496666666666666,0.779896556141616,0.6496666666666666,0.6931844557591907