A intro for train your model
Preprocess data
cd Gesture_Lexicon
python c
adjust the config ./Config/MOCCA/config_vae.json5 Attention: lxm_dim: 192, make sure as same as step 1.
cd Gesture_Generator
python train.py ./Config/MOCCA/config_vae.json5
revise below config
--checkpoint_path /root/project/Audio2Gesture/Gesture_Generator/Training/MOCCA/_RNN_20231211_152647/Checkpoints/trained_model.pth \
--checkpoint_config /root/project/Audio2Gesture/Gesture_Generator/Training/MOCCA/_RNN_20231211_152647/config.json5 \
--lxc_checkpoint_path /root/project/Audio2Gesture/Gesture_Lexicon/Training/MOCCA/_vqvae1d_20231206_115324/Checkpoints/trained_model.pth\
--lxc_checkpoint_config /root/project/Audio2Gesture/Gesture_Lexicon/Training/MOCCA/_vqvae1d_20231206_115324/config.json5 \
run ./train_gesture_vae.sh
Note: lexicon_size: 2048 in ./Data/MOCCA/Processed_4/Training_Data/config.json5
python train.py ./Config/MOCCA/config.json5