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

Issue with reproducing the performance of the model presented in the paper #8

Closed hwang9u closed 1 year ago

hwang9u commented 1 year ago

Hello. Thank you for your excellent work. I'm also very interested in speech emotion recognition!!

I am experiencing an issue with reproducing the performance of the model presented in the paper. Would it be possible for you to share the results of your experiments, such as the loss curves and the results on each dataset? I would like to compare my experiments with yours.

Thank you very much. Have a great day!!

Jiaxin-Ye commented 1 year ago

Thanks for your attention! The result and model of each corpus is available and you can train your own model to compare with them. However, for the loss curves, we didn't save all curves in our experiments and it is hard for us to find all the curve data again. We apologize for any inconvenience caused to you. Furthermore, due to hardware and software differences and randomness and initialization, the reproducing results may differ from our results. Considering this situation, we provide our pretrained models to let you test. If you think this comparison is unfair, you can retrain all the baselines in your local environment for comparative experiments.

Thanks.