Open feufhd opened 8 months ago
@feufhd 大哥 请问train.py中的导入包from animatediff.utils.util import save_videos_grid, pad_image这个代码 那个pad_image在这个util里面根本没有啊 请问你是怎么训练的呢 谢谢
@feufhd
According to the magic_animate paper, stage2 is designed to train temporal modules.
Thus, you need to initialize your model from the stage1, and then modify the aa_train_stage1.yaml, by making the trainable_modules.
trainable_modules:
- "motion_modules."
Thank you for your great work! I've finished training for stage 1, and I'd like to know how to launch the training for stage 2. Thank you!