aboutsummaryrefslogtreecommitdiffstats
path: root/modules/hypernetworks/ui.py
Commit message (Collapse)AuthorAgeFilesLines
* suggestions and fixes from the PRAUTOMATIC2023-05-101-2/+2
|
* ruff auto fixesAUTOMATIC2023-05-101-2/+2
|
* imports cleanup for ruffAUTOMATIC2023-05-101-2/+0
|
* extra networks UIAUTOMATIC2023-01-211-3/+2
| | | | rework of hypernets: rather than via settings, hypernets are added directly to prompt as <hypernet:name:weight>
* Variable dropout ratearia1th2023-01-101-2/+2
| | | | | | | | | | Implements variable dropout rate from #4549 Fixes hypernetwork multiplier being able to modified during training, also fixes user-errors by setting multiplier value to lower values for training. Changes function name to match torch.nn.module standard Fixes RNG reset issue when generating previews by restoring RNG state
* implement train apiVladimir Mandic2022-12-241-27/+4
|
* We have duplicate linear nowAngelBottomless2022-10-301-1/+1
|
* Re enable linearAngelBottomless2022-10-281-1/+1
|
* Disable unavailable or duplicate optionsAngelBottomless2022-10-271-1/+2
|
* Weight initialization and More activation funcAngelBottomless2022-10-261-1/+3
| | | | | | | | | | | | | | | | add weight init add weight init option in create_hypernetwork fstringify hypernet info save weight initialization info for further debugging fill bias with zero for He/Xavier initialize LayerNorm with Normal fix loading weight_init
* Merge branch 'master' of upstreamdiscus04342022-10-221-2/+6
|\
| * Sanitize hypernet name input.timntorres2022-10-211-0/+3
| |
| * Merge branch 'master' into training-help-textAUTOMATIC11112022-10-211-2/+3
| |\
| * | allow overwrite old hnDepFA2022-10-191-2/+3
| | |
* | | add dropoutdiscus04342022-10-221-5/+5
| |/ |/|
* | Merge branch 'AUTOMATIC1111:master' into masterdiscus04342022-10-201-1/+1
|\ \
| * | allow float sizes for hypernet's layer_structureAUTOMATIC2022-10-201-1/+1
| |/
* / updatediscus04342022-10-201-1/+2
|/
* enable to write layer structure of hn himselfdiscus04342022-10-191-0/+4
|
* layer options moves into create hnet uidiscus04342022-10-191-2/+7
|
* reports that training with medvram is possible.AUTOMATIC2022-10-111-1/+1
|
* add an option to unload models during hypernetwork training to save VRAMAUTOMATIC2022-10-111-1/+3
|
* produce error when training with medvram/lowvram enabledAUTOMATIC2022-10-111-0/+2
|
* add option to select hypernetwork modules when creatingAUTOMATIC2022-10-111-2/+2
|
* more renamesAUTOMATIC2022-10-111-2/+2
|
* rename hypernetwork dir to hypernetworks to prevent clash with an old ↵AUTOMATIC2022-10-111-0/+43
filename that people who use zip instead of git clone will have