Closed TDIT-haha closed 1 year ago
BEDLAM render pipeline generates images, it does not directly generate data for training in npz
format.
Please see https://github.com/pixelite1201/BEDLAM/blob/master/data_processing/ReadMe.md repo for details on how to prepare data for training.
I followed this readme and found that the content of the CSV file I generated myself was empty, which differs greatly from the CSV file provided by the author. As for this problem, I haven't found the reason? Do you have any ideas?
Closing. Please follow our Issue Tracker guidelines: Do not ask several unrelated questions on the same issue
First of all, thank you very much for your work. I created a composite image myself according to the README process. However, compared to the data published by the author, each image folder of the author will correspond to an NPZ file, such as 20221010 3-10 500 Batch01hand Zoom Subb D 6fps. tar will have a corresponding 20221010 3-10 500 Batch01hand Zoom Subb D 6fps. npz. But the data I synthesized lacks the corresponding npz file. How was this npz file generated?