search

visual-layer / fastdup

fastdup is a powerful free tool designed to rapidly extract valuable insights from your image & video datasets. Assisting you to increase your dataset images & labels quality and reduce your data operations costs at an unparalleled scale.
Other
1.52k stars 74 forks source link

[Feature Request]: Similarity.csv is not showing image names anymore, it shows index names. #255

Closed abulus closed 10 months ago

abulus commented 10 months ago

Feature Name

image names

Feature Description

Hi dear devs,

Currently I am using version 1.36

In older versions we would be able to see image names at least on similarity.csv but not anymore. It shows indexes.

If there is a hidden parameter that I can't see, please inform me.

Thank you for your works. image_2023-08-21_133236801

Contact Information [Optional]

No response

abulus commented 10 months ago

Actually, If we directly run like fastdup.run(input_dir="", work_dir="" like this it shows image file paths in similarity.csv.

But if we follow the instructions from documentation: We go as follow: fd = fastdup.create(work_dir, image_dir) fd.run() it shows index instead of image names.

dbickson commented 10 months ago

hi @abulus. In fastdup v1 you can do fd.similarity() to get the similar filenames. Please try it out and let us know if this works for you. 

import fastdup
fd = fastdup.create(input_dir=xxxx, work_dir=xxxx)
fd.run()
df = fd.similarity()

Thanks