Hi,
I am using this AVA Pretrained model from model zoo - https://dl.fbaipublicfiles.com/pyslowfast/model_zoo/ava/SLOWFAST_32x2_R101_50_50_v2.1.pkl
With some testing on my videos, I have came accross a conclusion that for different classes (among 80 classes my usecase is on few 7-10 classes) there is no exact one threshold point to decide wether a class to be present in predicted list or not. I hope if you knew about this thing, if you have the exact the threshold values for all 80 classes, could you please provide this information? It will be really a great help to me and many who are spending a lot of time on just adjusting thresholds.
Hi,
I am using this AVA Pretrained model from model zoo - https://dl.fbaipublicfiles.com/pyslowfast/model_zoo/ava/SLOWFAST_32x2_R101_50_50_v2.1.pkl
With some testing on my videos, I have came accross a conclusion that for different classes (among 80 classes my usecase is on few 7-10 classes) there is no exact one threshold point to decide wether a class to be present in predicted list or not. I hope if you knew about this thing, if you have the exact the threshold values for all 80 classes, could you please provide this information? It will be really a great help to me and many who are spending a lot of time on just adjusting thresholds.