Closed tomoki-oke closed 5 years ago
valid
refers to whether the ingredient detection has a relatively high probability of being correct. For instance, in the first entry, (makes about 4 cups)
is correctly marked as valid=False
since the detection model was uncertain about whether that was an ingredient. If you know for sure the ingredients, then valid=True
always
I'm making my own dataset for trying your model. I've been made layer1.json and layer2.json. But I wonder about about det_ingrs.json file. This file contains id, ingredients and valid. I suppose that a value valid show whether an ingredients corresponding on its index exists or not in vocab.txt. Is it correct? If not, please let me know what det_ingrs.json means and how I make det_ingrs.json.