Thank you very much for the great work! It is really amazing and makes our life easier. I am working on creating subset of laion5B given certain text query and came across some issues:
1) I firstly check this command clip-retrieval filter --query "cat" --output_folder "cat/" --indice_folder "indice_folder" . and I realize that I need to do clip inference and index before running filter. If I would like to download all the examples with 'cat' query from Laion5B dataset, do I need to download the whole dataset, then run inference, index and filter to extract the meta data of "cat"?
2) Then I check the simple_filter.ipynb file but confused about this "/home/rom1504/indices_paths.json". Where can we download this json file? does it include the clip index results of all the Laion5B dataset?
Hi,
Thank you very much for the great work! It is really amazing and makes our life easier. I am working on creating subset of laion5B given certain text query and came across some issues:
1) I firstly check this command
clip-retrieval filter --query "cat" --output_folder "cat/" --indice_folder "indice_folder"
. and I realize that I need to do clip inference and index before running filter. If I would like to download all the examples with 'cat' query from Laion5B dataset, do I need to download the whole dataset, then run inference, index and filter to extract the meta data of "cat"?2) Then I check the simple_filter.ipynb file but confused about this
"/home/rom1504/indices_paths.json"
. Where can we download this json file? does it include the clip index results of all the Laion5B dataset?I appreciate your time and help!