search

ai4r / Gesture-Generation-from-Trimodal-Context

Speech Gesture Generation from the Trimodal Context of Text, Audio, and Speaker Identity (SIGGRAPH Asia 2020)
Other
245 stars 35 forks source link

text modality ablation study #43

Closed hslr4 closed 2 years ago

hslr4 commented 2 years ago

Hi, thank you for the great paper and the code. Could you please provide some code of the ablation study regarding the text modality? I'm especially interested in which speaker ID did you use for the gestures for synthesized speech and how is the replacement of words with WordNet implemented?

youngwoo-yoon commented 2 years ago

Hello, sorry for the late reply. I share the code here that I used for the text-altering experiment. This is not runnable, but you should be able to get the details. In this experiment, the speaker IDs were selected randomly for each sentence sample as shown in line 79 of the code.

hslr4 commented 2 years ago

Thank you 👍