danielcccc
@engrchrishenry The author just trains the model once over whole training set. The batch data is provided by training list. You can change the condition in the training loop and...
same issue， waiting for fixing
Same problem after merging CogVideo1.5X branch. It seems that this update cause the prolems:https://github.com/THUDM/CogVideo/commit/3a9af5bdd937faaa7803914aa94e384d7f40af67#diff-7b0a094155a48dc8761cbfd20bdf64fe0cc9021873b7ab23cea5a5cae38a670e