logical-reasoning / results /mgtv-llama3_p2_full_metrics.csv
dh-mc's picture
analysis of internlm/llama3
bc127a6
raw
history blame
487 Bytes
epoch,model,accuracy,precision,recall,f1
0,shenzhi-wang/Llama3-8B-Chinese-Chat,0.73,0.7709739363586101,0.73,0.7462914191370829
1,shenzhi-wang/Llama3-8B-Chinese-Chat_checkpoint-175,0.773,0.7739158621170704,0.773,0.7642801051494378
2,shenzhi-wang/Llama3-8B-Chinese-Chat_checkpoint-350,0.7046666666666667,0.814516278555831,0.7046666666666667,0.7453647242165446
3,shenzhi-wang/Llama3-8B-Chinese-Chat_checkpoint-525,0.6793333333333333,0.8030704466494853,0.6793333333333333,0.7246368106499855