aboutsummaryrefslogtreecommitdiffstats
path: root/extensions-builtin/Lora/network_oft.py
Commit message (Collapse)AuthorAgeFilesLines
* Merge pull request #14871 from v0xie/boftAUTOMATIC11112024-02-191-10/+48
| | | Support inference with LyCORIS BOFT networks
* fix: wrong devices for eye and constraintv0xie2024-01-221-2/+2
|
* Fix dtype casting for OFT moduleKohaku-Blueleaf2024-01-051-3/+3
|
* better naming and correct order for device.Kohaku-Blueleaf2023-12-131-2/+3
|
* remove debug printKohaku-Blueleaf2023-12-131-1/+0
|
* Use self.scale instead of custom finalizeKohaku-Blueleaf2023-12-131-18/+2
|
* Fix network_oftKohaku-Blueleaf2023-12-131-10/+11
|
* feat: LyCORIS/kohya OFT network supportv0xie2023-11-161-82/+26
|
* fix: ignore calc_scale() for COFT which has very small alphav0xie2023-11-151-11/+5
|
* refactor: remove unused functionv0xie2023-11-041-47/+0
|
* refactor: use same updown for both kohya OFT and LyCORIS diag-oftv0xie2023-11-041-17/+74
|
* refactor: move factorization to lyco_helpers, separate calc_updown for kohya ↵v0xie2023-11-041-101/+30
| | | | and kb
* skip multihead attn for nowv0xie2023-11-041-17/+37
|
* no idea what i'm doing, trying to support both type of OFT, kblueleaf ↵v0xie2023-11-021-47/+145
| | | | diag_oft has MultiheadAttn which kohya's doesn't?, attempt create new module based off network_lora.py, errors about tensor dim mismatch
* test implementation based on kohaku diag-oft implementationv0xie2023-11-021-21/+38
|
* style: conform stylev0xie2023-10-221-1/+1
|
* fix: multiplier applied twice in finalize_updownv0xie2023-10-221-1/+22
|
* refactor: remove used OFT functionsv0xie2023-10-221-72/+10
|
* fix: use merge_weight to cache valuev0xie2023-10-221-17/+40
|
* style: cleanup oftv0xie2023-10-211-75/+7
|
* fix: support multiplier, no forward pass hookv0xie2023-10-211-10/+33
|
* fix: return orig weights during updown, merge weights before forwardv0xie2023-10-211-21/+69
|
* refactor: use forward hook instead of custom forwardv0xie2023-10-211-9/+24
|
* style: fix ambiguous variable namev0xie2023-10-191-2/+2
|
* style: formattingv0xie2023-10-191-2/+2
|
* refactor: fix constraint, re-use get_weightv0xie2023-10-191-24/+16
|
* faster by calculating R in updown and using cached R in forwardv0xie2023-10-181-7/+8
|
* faster by using cached R in forwardv0xie2023-10-181-3/+14
|
* inference working but SLOWv0xie2023-10-181-37/+36
|
* wip incorrect OFT implementationv0xie2023-10-181-0/+82
generated by cgit v1.2.3 (git 2.25.1) at 2025-04-28 22:31:52 +0000