Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | Add FP32 fallback support on sd_vae_approx | hidenorly | 2023-11-20 | 1 | -1/+7 |
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This tries to execute interpolate with FP32 if it failed. Background is that on some environment such as Mx chip MacOS devices, we get error as follows: ``` "torch/nn/functional.py", line 3931, in interpolate return torch._C._nn.upsample_nearest2d(input, output_size, scale_factors) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ RuntimeError: "upsample_nearest2d_channels_last" not implemented for 'Half' ``` In this case, ```--no-half``` doesn't help to solve. Therefore this commits add the FP32 fallback execution to solve it. Note that the submodule may require additional modifications. The following is the example modification on the other submodule. ```repositories/stable-diffusion-stability-ai/ldm/modules/diffusionmodules/openaimodel.py class Upsample(nn.Module): ..snip.. def forward(self, x): assert x.shape[1] == self.channels if self.dims == 3: x = F.interpolate( x, (x.shape[2], x.shape[3] * 2, x.shape[4] * 2), mode="nearest" ) else: try: x = F.interpolate(x, scale_factor=2, mode="nearest") except: x = F.interpolate(x.to(th.float32), scale_factor=2, mode="nearest").to(x.dtype) if self.use_conv: x = self.conv(x) return x ..snip.. ``` You can see the FP32 fallback execution as same as sd_vae_approx.py. | ||||
* | Merge pull request #14009 from ↵ | AUTOMATIC1111 | 2023-11-20 | 2 | -4/+21 |
|\ | | | | | | | | | AUTOMATIC1111/Option-to-show-batch-img2img-results-in-UI Option to show batch img2img results in UI | ||||
| * | Option to show batch img2img results in UI | w-e-w | 2023-11-19 | 2 | -4/+21 |
| | | | | | | | | | | | | | | | | shared.opts.img2img_batch_show_results_limit limit the number of images return to the UI for batch img2img default limit 32 0 no images are shown -1 unlimited, all images are shown | ||||
* | | Merge branch 'dag' into dev | AUTOMATIC1111 | 2023-11-20 | 2 | -153/+148 |
|\ \ | |||||
| * | | rework extensions metadata: use custom sorter that doesn't mess the order as ↵ | AUTOMATIC1111 | 2023-11-20 | 2 | -153/+148 |
| | | | | | | | | | | | | much and ignores cyclic errors, use classes with named fields instead of dictionaries, eliminate some duplicated code | ||||
* | | | Merge pull request #13944 from wfjsw/dag | AUTOMATIC1111 | 2023-11-20 | 2 | -24/+188 |
|\| | | | | | | | | implementing script metadata and DAG sorting mechanism | ||||
| * | | use metadata.ini for meta filename | wfjsw | 2023-11-19 | 1 | -6/+6 |
| | | | |||||
| * | | bug fix | wfjsw | 2023-11-11 | 1 | -7/+18 |
| | | | |||||
| * | | fix | wfjsw | 2023-11-11 | 1 | -1/+0 |
| | | | |||||
| * | | allow comma and whitespace as separator | wfjsw | 2023-11-11 | 2 | -6/+9 |
| | | | |||||
| * | | remove the assumption of same name | wfjsw | 2023-11-11 | 1 | -51/+30 |
| | | | |||||
| * | | reverse the extension load order so builtin extensions load earlier natively | wfjsw | 2023-11-11 | 1 | -1/+1 |
| | | | |||||
| * | | populate loaded_extensions from extension list instead | wfjsw | 2023-11-11 | 1 | -3/+2 |
| | | | |||||
| * | | implementing script metadata and DAG sorting mechanism | wfjsw | 2023-11-11 | 2 | -24/+197 |
| |/ | |||||
* | | Merge pull request #14035 from AUTOMATIC1111/sysinfo-json | AUTOMATIC1111 | 2023-11-20 | 2 | -2/+2 |
|\ \ | | | | | | | save sysinfo as .json | ||||
| * | | save sysinfo as .json | w-e-w | 2023-11-19 | 2 | -2/+2 |
|/ / | | | | | | | GitHub now allows uploading of .json files in issues | ||||
* | | Merge pull request #13931 from AUTOMATIC1111/style-hotkeys | AUTOMATIC1111 | 2023-11-19 | 1 | -2/+2 |
|\ \ | | | | | | | Enable prompt hotkeys in style editor | ||||
| * | | Enable prompt hotkeys in style editor | missionfloyd | 2023-11-10 | 1 | -2/+2 |
| | | | |||||
* | | | Merge pull request #13014 from ↵ | AUTOMATIC1111 | 2023-11-19 | 4 | -8/+30 |
|\ \ \ | | | | | | | | | | | | | | | | | AUTOMATIC1111/thread-safe-extranetworks-list_items thread safe extra network list_items | ||||
| * | | | add comment | w-e-w | 2023-09-12 | 4 | -0/+4 |
| | | | | |||||
| * | | | return nothing if not found | w-e-w | 2023-09-11 | 4 | -5/+20 |
| | | | | |||||
| * | | | thread safe extra network using list | w-e-w | 2023-09-09 | 3 | -3/+6 |
| | | | | |||||
| * | | | Revert "thread safe extra network list_items" | w-e-w | 2023-09-09 | 5 | -16/+12 |
| | | | | | | | | | | | | | | | | This reverts commit aab385d01b4311726127397552d791f4d71b7147. | ||||
| * | | | thread safe extra network list_items | w-e-w | 2023-09-03 | 5 | -12/+16 |
| | | | | |||||
* | | | | Merge pull request #13929 from kingljl/fix-dependency-address-patch-1 | AUTOMATIC1111 | 2023-11-19 | 1 | -1/+1 |
|\ \ \ \ | | | | | | | | | | | Fix dependency address patch 1 | ||||
| * | | | | Update README.md | fuchen.ljl | 2023-11-10 | 1 | -1/+1 |
| | | | | | | | | | | | | | | | Modify the stablediffusion dependency address | ||||
| * | | | | Update README.md | fuchen.ljl | 2023-11-10 | 0 | -0/+0 |
| | |/ / | |/| | | | | | | Modify the stablediffusion dependency address | ||||
* | | | | Merge pull request #13962 from kaalibro/dev | AUTOMATIC1111 | 2023-11-19 | 1 | -3/+12 |
|\ \ \ \ | | | | | | | | | | | Fixes generation restart not working for some users when 'Ctrl+Enter' is pressed | ||||
| * | | | | Fixes generation restart not working for some users when 'Ctrl+Enter' is pressed | kaalibro | 2023-11-13 | 1 | -3/+12 |
| | |_|/ | |/| | | |||||
* | | | | Merge pull request #13968 from kaalibro/extranetworks-path-sorting | AUTOMATIC1111 | 2023-11-19 | 2 | -2/+3 |
|\ \ \ \ | | | | | | | | | | | Adds 'Path' sorting for Extra network cards | ||||
| * | | | | Adds 'Path' sorting for Extra network cards | kaalibro | 2023-11-13 | 2 | -2/+3 |
| |/ / / | |||||
* | | | | Merge pull request #13977 from AUTOMATIC1111/hotfix-postprocessing-state-end | AUTOMATIC1111 | 2023-11-19 | 1 | -1/+1 |
|\ \ \ \ | | | | | | | | | | | Hotfix: call shared.state.end() after postprocessing done | ||||
| * | | | | hotfix: call shared.state.end() after postprocessing done | AngelBottomless | 2023-11-15 | 1 | -1/+1 |
| |/ / / | |||||
* | | | | Merge pull request #13692 from v0xie/network-oft | AUTOMATIC1111 | 2023-11-19 | 3 | -0/+157 |
|\ \ \ \ | | | | | | | | | | | Support inference with OFT networks | ||||
| * | | | | feat: LyCORIS/kohya OFT network support | v0xie | 2023-11-16 | 1 | -82/+26 |
| | | | | | |||||
| * | | | | fix: ignore calc_scale() for COFT which has very small alpha | v0xie | 2023-11-15 | 1 | -11/+5 |
| | | | | | |||||
| * | | | | Merge pull request #2 from v0xie/network-oft-change-impl | v0xie | 2023-11-04 | 1 | -17/+27 |
| |\ \ \ \ | | | | | | | | | | | | | Use same updown implementation for LyCORIS OFT as kohya-ss OFT | ||||
| | * | | | | refactor: remove unused function | v0xie | 2023-11-04 | 1 | -47/+0 |
| | | | | | | |||||
| | * | | | | refactor: use same updown for both kohya OFT and LyCORIS diag-oft | v0xie | 2023-11-04 | 1 | -17/+74 |
| |/ / / / | |||||
| * | | | | Merge pull request #1 from v0xie/oft-faster | v0xie | 2023-11-04 | 3 | -17/+135 |
| |\ \ \ \ | | | | | | | | | | | | | | | | | | | | | | | | | Support LyCORIS diag-oft OFT implementation (minus MultiheadAttention layer), maintains support for kohya-ss OFT | ||||
| | * | | | | refactor: move factorization to lyco_helpers, separate calc_updown for kohya ↵ | v0xie | 2023-11-04 | 2 | -101/+77 |
| | | | | | | | | | | | | | | | | | | | | | | | | and kb | ||||
| | * | | | | skip multihead attn for now | v0xie | 2023-11-04 | 1 | -17/+37 |
| | | | | | | |||||
| | * | | | | no idea what i'm doing, trying to support both type of OFT, kblueleaf ↵ | v0xie | 2023-11-02 | 1 | -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 | ||||
| | * | | | | detect diag_oft type | v0xie | 2023-11-02 | 1 | -0/+7 |
| | | | | | | |||||
| | * | | | | test implementation based on kohaku diag-oft implementation | v0xie | 2023-11-02 | 1 | -21/+38 |
| |/ / / / | |||||
| * | | | | style: conform style | v0xie | 2023-10-22 | 1 | -1/+1 |
| | | | | | |||||
| * | | | | fix: multiplier applied twice in finalize_updown | v0xie | 2023-10-22 | 1 | -1/+22 |
| | | | | | |||||
| * | | | | refactor: remove used OFT functions | v0xie | 2023-10-22 | 1 | -72/+10 |
| | | | | | |||||
| * | | | | fix: use merge_weight to cache value | v0xie | 2023-10-22 | 1 | -17/+40 |
| | | | | | |||||
| * | | | | style: cleanup oft | v0xie | 2023-10-21 | 1 | -75/+7 |
| | | | | |