16 Commits

Author	SHA1	Message	Date
Kohaku-Blueleaf	2100f4cac2	linting	2025-07-23 14:37:30 +08:00
Kohaku-Blueleaf	dc05edb5b2	Add OFT training	2025-07-23 12:38:47 +08:00
Kohaku-Blueleaf	c92c884fec	use more stable init	2025-07-23 12:11:24 +08:00
Kohaku-Blueleaf	7a677763e8	Fix bugs of loha	2025-07-23 11:55:48 +08:00
Kohaku-Blueleaf	28a236d505	Use custom fwd/bwd func and better init for loha	2025-07-23 11:31:09 +08:00
Kohaku-Blueleaf	c025308918	correct key name for loha	2025-07-22 16:34:11 +08:00
Kohaku-Blueleaf	bf053a9c55	Update __init__.py	2025-07-22 16:33:45 +08:00
Kohaku-Blueleaf	7aa3184e85	Add adapter map for algo selection	2025-07-22 16:29:32 +08:00
Kohaku-Blueleaf	f5bdfcf0fe	Add loha train impl	2025-07-22 16:29:24 +08:00
Kohaku-Blueleaf	2345fbff3e	Add lokr train impl	2025-07-22 16:29:10 +08:00
Kohaku-Blueleaf	046c517a1f	Add factorization utils for lokr	2025-07-22 16:29:04 +08:00
Kohaku-Blueleaf	520eb77b72	LoRA Trainer: LoRA training node in weight adapter scheme (#8446)	2025-06-13 19:25:59 -04:00
Kohaku-Blueleaf	2ab9618732	Fix the bugs in OFT/BOFT moule (#7909) ... * Correct calculate_weight and load for OFT * Correct calculate_weight and loading for BOFT	2025-05-02 13:12:37 -04:00
Kohaku-Blueleaf	a8f63c0d5b	Support dora_scale on both axis (#7727)	2025-04-22 05:01:27 -04:00
Kohaku-Blueleaf	966c43ce26	Add OFT/BOFT algorithm in weight adapter (#7725)	2025-04-22 04:59:47 -04:00
Kohaku-Blueleaf	1f3fba2af5	Unified Weight Adapter system for better maintainability and future feature of Lora system (#7540)	2025-04-21 20:15:32 -04:00