lixiangucas01 / GLAM

This is the official code for paper "Speech Emotion Recognition with Global-Aware Fusion on Multi-scale Feature Representation" published in ICASSP 2022
43 stars 10 forks source link

Reproduce problem #1

Open SteveTanggithub opened 2 years ago

SteveTanggithub commented 2 years ago

hello, we cannot reproduce the results by running GLAM and GLAM_all in ur paper. Can you provide more information about the training parameters?

Thinkre commented 2 years ago

default parameters are all set in train.py. For different settings, you can just run bash script in script dir.