summaryrefslogtreecommitdiffstats
path: root/train_ti.py
Commit message (Collapse)AuthorAgeFilesLines
* Support LoRA training for token embeddingsVolpeon2023-05-161-6/+4
|
* UpdateVolpeon2023-05-111-0/+24
|
* UpdateVolpeon2023-05-061-0/+17
|
* UpdateVolpeon2023-05-051-13/+32
|
* UpdateVolpeon2023-04-271-2/+11
|
* Fix cycle loopVolpeon2023-04-231-3/+5
|
* UpdateVolpeon2023-04-201-4/+9
|
* FixVolpeon2023-04-171-3/+3
|
* Improved automation capsVolpeon2023-04-161-20/+33
|
* Added option to use constant LR on cycles > 1Volpeon2023-04-161-0/+9
|
* FixVolpeon2023-04-161-14/+7
|
* UpdateVolpeon2023-04-161-1/+5
|
* TI via LoRAVolpeon2023-04-151-6/+22
|
* Added cycle LR decayVolpeon2023-04-131-3/+5
|
* UpdateVolpeon2023-04-131-3/+10
|
* Experimental convnext discriminator supportVolpeon2023-04-111-0/+14
|
* UpdateVolpeon2023-04-111-7/+16
|
* Randomize dataset across cyclesVolpeon2023-04-101-2/+4
|
* UpdateVolpeon2023-04-101-27/+41
|
* UpdateVolpeon2023-04-091-3/+14
|
* FixVolpeon2023-04-091-1/+1
|
* UpdateVolpeon2023-04-091-24/+42
|
* UpdateVolpeon2023-04-081-5/+18
|
* UpdateVolpeon2023-04-081-7/+5
|
* Run PTI only if placeholder tokens arg isn't emptyVolpeon2023-04-071-6/+6
|
* UpdateVolpeon2023-04-071-3/+3
|
* Add color jitterVolpeon2023-04-051-1/+12
|
* Fix choice argsVolpeon2023-04-041-8/+9
|
* TI: Bring back old embedding decayVolpeon2023-04-041-5/+19
|
* Improved sparse embeddingsVolpeon2023-04-031-1/+1
|
* TI: Delta learningVolpeon2023-04-031-26/+11
|
* TI: No tag dropout by defaultVolpeon2023-04-031-1/+1
|
* Bring back Lion optimizerVolpeon2023-04-031-3/+27
|
* Update dataset format: Separate prompt and keywordsVolpeon2023-04-021-1/+1
|
* RevertVolpeon2023-04-011-6/+46
|
* Combined TI with embedding and LoRAVolpeon2023-04-011-25/+5
|
* Experimental: TI via LoRAVolpeon2023-04-011-22/+2
|
* UpdateVolpeon2023-04-011-1/+0
|
* Add support for Adafactor, add TI initializer noiseVolpeon2023-04-011-2/+23
|
* UpdateVolpeon2023-03-311-1/+3
|
* UpdateVolpeon2023-03-311-0/+7
|
* FixVolpeon2023-03-311-1/+2
|
* FixVolpeon2023-03-311-1/+1
|
* Support Dadaptation d0, adjust sample freq when steps instead of epochs are usedVolpeon2023-03-311-4/+11
|
* FixVolpeon2023-03-311-1/+2
|
* FixVolpeon2023-03-281-1/+1
|
* Support num_train_steps arg againVolpeon2023-03-281-9/+21
|
* Fix TIVolpeon2023-03-271-8/+8
|
* Fix TIVolpeon2023-03-271-1/+10
|
* Fix TIVolpeon2023-03-271-1/+1
|