8 Commits

Author	SHA1	Message	Date
OleehyO	455b44a7b5	chore: code cleanup and parameter optimization ... - Remove redundant comments and debug information - Adjust default parameters in training scripts - Clean up code in lora_trainer and trainer implementations	2025-01-13 11:56:28 +00:00
zR	1534bf33eb	add pipeline	2025-01-12 19:27:21 +08:00
OleehyO	e213b6c083	fix: pad latent frames to match patch_size_t requirements	2025-01-11 02:08:07 +00:00
OleehyO	36427274d6	style: format import statements across finetune module	2025-01-07 05:54:52 +00:00
zR	1789f07256	format and check fp16 for cogvideox2b	2025-01-07 13:16:18 +08:00
OleehyO	9157e0cbc8	Adapt dataset for text embeddings and add noise padding ... - Add text embedding support in dataset collation - Pad 2 random noise frames at the beginning of latent space during training	2025-01-06 10:44:58 +00:00
OleehyO	7e1ac76847	feat(cogvideox): add prompt embedding caching support ... This change enables caching of prompt embeddings in the CogVideoX text-to-video LoRA trainer, which can improve training efficiency by avoiding redundant text encoding operations.	2025-01-04 06:17:56 +00:00
OleehyO	a001842834	feat: implement CogVideoX trainers for I2V and T2V tasks ... Add and refactor trainers for CogVideoX model variants: - Implement CogVideoXT2VLoraTrainer for text-to-video generation - Refactor CogVideoXI2VLoraTrainer for image-to-video generation Both trainers support LoRA fine-tuning with proper handling of: - Model components loading and initialization - Video encoding and batch collation - Loss computation with noise prediction - Validation step for generation	2025-01-01 15:10:54 +00:00