Hello, thank you very much for your work and open source!
I want to use the pre-trained weights on your AffectNet data set, but I found that the model can only predict the probability distribution of seven categories, but cannot determine the specific expression corresponding to each dimension of the pre-trained model, such as outputting a certain image tensor( [[ 1.9637, -1.2740, 2.0596, 0.2915, 0.2487, -3.2745, -0.2642]], but I am not sure what emotion the 2.02596 position corresponds to.
Hello, thank you very much for your work and open source! I want to use the pre-trained weights on your AffectNet data set, but I found that the model can only predict the probability distribution of seven categories, but cannot determine the specific expression corresponding to each dimension of the pre-trained model, such as outputting a certain image tensor( [[ 1.9637, -1.2740, 2.0596, 0.2915, 0.2487, -3.2745, -0.2642]], but I am not sure what emotion the 2.02596 position corresponds to.
Can the author provide it? Thank you so much!