Hi, thanks for the open source! I have several questions regarding dataset processing, model training and evaluation.
-
It seems that EMOCA feature and HuBERT features are required in python vico_preprocessing.py, could you please share details on how to get these features?
-
Are the four steps of training required in their given order before I'm able get the final visualization results? Could you share the pre-trained weights of the model?
-
The model evaluation codes seems to only give the numerical results, how can I get the visualization results of vico datasets like you demonstrated in the project page?
Hi, thanks for the open source! I have several questions regarding dataset processing, model training and evaluation.
It seems that EMOCA feature and HuBERT features are required in
python vico_preprocessing.py, could you please share details on how to get these features?Are the four steps of training required in their given order before I'm able get the final visualization results? Could you share the pre-trained weights of the model?
The model evaluation codes seems to only give the numerical results, how can I get the visualization results of vico datasets like you demonstrated in the project page?