search

Jiaxin-Ye / TIM-Net_SER

[ICASSP 2023] Official Tensorflow implementation of "Temporal Modeling Matters: A Novel Temporal Emotional Modeling Approach for Speech Emotion Recognition".
GNU General Public License v3.0
161 stars 25 forks source link

Equation (3) of paper is inconsistent with Code #3

Closed vtsouval closed 1 year ago

vtsouval commented 1 year ago

Hello and thank you for open sourcing your code. I have noticed an inconsistency between Fig.1 and Equation 3 of your paper. Specifically, the outputs of the two TAB blocks are added and then passed though the GlobalAveragePooling1D layer in your code (and Figure), while in Equation 3 you describe them in the reverse order. Can you clarify this for me? Thank you again for your vital contribution to the research community.

Jiaxin-Ye commented 1 year ago

Hi! Thank you for indicating the issue. Although the output obtained from both calculations is consistent, the formulation of the paper is not well presented. We have fixed this issue in our latest version (camera-ready version). I think that answers your question.

Thanks again.