generating multiple duplicate captions in ImageCaptioning

[rakashi]

Hi,
In "IMAGE CAPTIONING", we have given beamsize as 3,as it generating 3 captions for that particular image, but it is generating duplicate captions twice or thrice with different p value,

In the below output 0,2 are same captions with different p value..

can anyone help me out how to avoid duplicate captions and to get unique captions for particular image..

OUTPUT:
INFO:tensorflow:Successfully loaded checkpoint: model.ckpt-102000
Captions for image = 3.jpg
0) a group of people standing around a table (p=0.019730)
1) a group of people standing in a room . (p=0.026106)
2) a group of people standing around a table . (p=0.028524)