Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | Added CLI arg to set dataloader worker num; improved text encoder handling ↵ | Volpeon | 2022-10-27 | 3 | -11/+50 |
| | | | | with Dreambooth | ||||
* | Euler_a: Re-introduce generator arg for reproducible output | Volpeon | 2022-10-27 | 2 | -3/+10 |
| | |||||
* | New Euler_a scheduler | Volpeon | 2022-10-26 | 6 | -318/+215 |
| | |||||
* | Advanced datasets | Volpeon | 2022-10-24 | 3 | -75/+115 |
| | |||||
* | Update | Volpeon | 2022-10-24 | 6 | -17/+415 |
| | |||||
* | Revert lat; fix skip attribute in dataset | Volpeon | 2022-10-22 | 2 | -76/+87 |
| | |||||
* | Training update | Volpeon | 2022-10-22 | 2 | -87/+75 |
| | |||||
* | Add optional TI functionality to Dreambooth | Volpeon | 2022-10-21 | 3 | -45/+77 |
| | |||||
* | Dreambooth: Added option to insert a new input token; removed Dreambooth Plus | Volpeon | 2022-10-19 | 2 | -1029/+42 |
| | |||||
* | Updated Dreambooth training | Volpeon | 2022-10-19 | 3 | -37/+76 |
| | |||||
* | Adapted other scripts for new prompt processing | Volpeon | 2022-10-18 | 3 | -17/+23 |
| | |||||
* | Improved prompt handling | Volpeon | 2022-10-18 | 5 | -129/+100 |
| | |||||
* | Implemented extended prompt limit | Volpeon | 2022-10-17 | 4 | -58/+108 |
| | |||||
* | Upstream updates; better handling of textual embedding | Volpeon | 2022-10-17 | 5 | -59/+82 |
| | |||||
* | Update | Volpeon | 2022-10-16 | 3 | -9/+15 |
| | |||||
* | Update | Volpeon | 2022-10-16 | 3 | -33/+69 |
| | |||||
* | Update | Volpeon | 2022-10-15 | 3 | -26/+49 |
| | |||||
* | Removed aesthetic gradients; training improvements | Volpeon | 2022-10-14 | 6 | -243/+77 |
| | |||||
* | Added support for Aesthetic Gradients | Volpeon | 2022-10-13 | 8 | -61/+245 |
| | |||||
* | Added TI+Dreambooth training | Volpeon | 2022-10-13 | 4 | -12/+961 |
| | |||||
* | Various updates | Volpeon | 2022-10-12 | 7 | -223/+142 |
| | |||||
* | Added EMA support to Textual Inversion | Volpeon | 2022-10-10 | 2 | -17/+60 |
| | |||||
* | Remove unused code | Volpeon | 2022-10-10 | 1 | -138/+0 |
| | |||||
* | TI: Use different lr scheduler | Volpeon | 2022-10-10 | 1 | -1/+1 |
| | |||||
* | Dreambooth: Add EMA support | Volpeon | 2022-10-10 | 1 | -6/+44 |
| | |||||
* | Updated default params | Volpeon | 2022-10-10 | 2 | -7/+7 |
| | |||||
* | Update | Volpeon | 2022-10-09 | 4 | -13/+146 |
| | |||||
* | Update | Volpeon | 2022-10-08 | 6 | -144/+169 |
| | |||||
* | Fix small details | Volpeon | 2022-10-07 | 3 | -57/+59 |
| | |||||
* | Training: Create multiple class images per training image | Volpeon | 2022-10-07 | 4 | -34/+101 |
| | |||||
* | Update | Volpeon | 2022-10-06 | 8 | -1116/+159 |
| | |||||
* | Inference: Add support for embeddings | Volpeon | 2022-10-04 | 1 | -2/+30 |
| | |||||
* | Multi-vector TI was broken | Volpeon | 2022-10-04 | 1 | -69/+38 |
| | |||||
* | Add Textual Inversion with class dataset (a la Dreambooth) | Volpeon | 2022-10-04 | 5 | -47/+968 |
| | |||||
* | Default sample steps 30 -> 40 | Volpeon | 2022-10-04 | 2 | -2/+2 |
| | |||||
* | Fix: Prompts | Volpeon | 2022-10-04 | 1 | -2/+6 |
| | |||||
* | Bugfix | Volpeon | 2022-10-04 | 2 | -6/+6 |
| | |||||
* | Fix | Volpeon | 2022-10-03 | 1 | -8/+2 |
| | |||||
* | Dreambooth: Generate specialized class images from input prompts | Volpeon | 2022-10-03 | 5 | -164/+129 |
| | |||||
* | Small perf improvements | Volpeon | 2022-10-03 | 5 | -51/+58 |
| | |||||
* | Added negative prompt support for training scripts | Volpeon | 2022-10-03 | 4 | -133/+63 |
| | |||||
* | Fixed euler_a generator argument | Volpeon | 2022-10-03 | 5 | -23/+10 |
| | |||||
* | Assign unused images in validation dataset to train dataset | Volpeon | 2022-10-03 | 4 | -9/+26 |
| | |||||
* | Added script to convert Differs -> SD | Volpeon | 2022-10-03 | 1 | -0/+234 |
| | |||||
* | Use euler_a for samples in learning scripts; backported improvement from ↵ | Volpeon | 2022-10-03 | 4 | -203/+230 |
| | | | | Dreambooth to Textual Inversion | ||||
* | Small performance improvements | Volpeon | 2022-10-02 | 3 | -17/+28 |
| | |||||
* | Fix img2img | Volpeon | 2022-10-02 | 3 | -60/+38 |
| | |||||
* | More consistent euler_a | Volpeon | 2022-10-02 | 3 | -45/+33 |
| | |||||
* | WIP: img2img | Volpeon | 2022-10-02 | 2 | -39/+131 |
| | |||||
* | Fix seed, better progress bar, fix euler_a for batch size > 1 | Volpeon | 2022-10-01 | 3 | -15/+15 |
| |