请问运行结果准确率怎么看？

[yxj1992]

`Run id: X2GQ7H Log directory: /tmp/tflearn_logs/

Training samples: 25 Validation samples: 17

Training Step: 1 | time: 4.402s | Adam | epoch: 001 | loss: 0.00000 - acc: 0.0000 | val_loss: 0.69363 - val_acc: 0.4118 -- iter: 25/25

[[0.49861583 0.50138414] [0.49942857 0.50057137] [0.4990096 0.50099045] [0.49668297 0.503317 ] [0.5001381 0.4998619 ] [0.49681374 0.50318635] [0.49952838 0.5004716 ] [0.4993792 0.50062084] [0.49945635 0.5005436 ] [0.49686605 0.50313395] [0.49935415 0.5006458 ] [0.49935263 0.5006474 ] [0.49893185 0.5010681 ] [0.49781716 0.5021828 ] [0.49845394 0.5015461 ] [0.49686605 0.503134 ] [0.4986591 0.50134087] [0.49832577 0.5016743 ] [0.49850422 0.5014958 ] [0.49901885 0.50098115] [0.49669054 0.5033094 ] [0.49686566 0.5031344 ] [0.4995494 0.5004506 ] [0.49870756 0.5012924 ] [0.4988313 0.5011687 ]] `

[jimenbian]

不是两列嘛，每一行对应一个句子，左边是0的权重，右边是1的权重。0是负面情绪，1是正面。

[yxj1992]

但是不知道每一行这个句子本来是正面还是负面，也就是说没有一个总的准确率。希望您能加上，这样实验结果更完善一些