Hi,
If you have a CSV of bounding box labels such that one column is ID->IMG_1705.jpeg, IMG_1706.jpeg …etc. the other column is Label->[540,697,1122,946], [w, x, y, z], etc:
How should I go about creating a data bunch?
Some thoughts:
- It looks like “get_annotations” is the factory method, but I don’t have my labels in JSON format to work with the function. Should I try to force my data into that format?
- I have the ImageBBox.create function working to create a BBox based on my data (hardcoded), but I can’t make the leap from a single example to a data bunch.
- Its also not clear to me at what point the BBox’s would be created if I used “Step 3: Label the Inputs” from the data_block docs either. It looks like “Label_from_func” or “Label_from_df” are the only ones that could possibly work, but how would they know it is an ImageBBox class?
below should be the image of my ImageBBox.create function:
Much thanks