Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | Fix forced reload | Kohaku-Blueleaf | 2023-12-06 | 1 | -1/+1 |
| | |||||
* | Ensure the cached weight will not be affected | Kohaku-Blueleaf | 2023-12-02 | 1 | -2/+2 |
| | |||||
* | Merge branch 'dev' into test-fp8 | Kohaku-Blueleaf | 2023-12-02 | 1 | -4/+13 |
|\ | |||||
| * | Add support for SD 2.1 Turbo, by converting the state dict from SGM to LDM ↵ | MrCheeze | 2023-12-02 | 1 | -4/+13 |
| | | | | | | | | on load | ||||
* | | Fix pre-fp8 | Kohaku-Blueleaf | 2023-11-25 | 1 | -1/+1 |
| | | |||||
* | | Option for using fp16 weight when apply lora | Kohaku-Blueleaf | 2023-11-21 | 1 | -3/+11 |
| | | |||||
* | | Use options instead of cmd_args | Kohaku-Blueleaf | 2023-11-19 | 1 | -29/+32 |
| | | |||||
* | | Merge branch 'dev' into test-fp8 | Kohaku-Blueleaf | 2023-11-16 | 1 | -1/+4 |
|\| | |||||
| * | more changes for #13865: fix formatting, rename the function, add comment ↵ | AUTOMATIC1111 | 2023-11-05 | 1 | -1/+1 |
| | | | | | | | | and add a readme entry | ||||
| * | linter | AUTOMATIC1111 | 2023-11-05 | 1 | -2/+2 |
| | | |||||
| * | Merge branch 'dev' into master | AUTOMATIC1111 | 2023-11-05 | 1 | -20/+30 |
| |\ | |||||
| * | | Use devices.torch_gc() instead of empty_cache() | Ritesh Gangnani | 2023-11-05 | 1 | -1/+0 |
| | | | |||||
| * | | Add SSD-1B as a supported model | Ritesh Gangnani | 2023-11-05 | 1 | -2/+6 |
| | | | |||||
* | | | ManualCast for 10/16 series gpu | Kohaku-Blueleaf | 2023-10-28 | 1 | -9/+12 |
| | | | |||||
* | | | ignore mps for fp8 | Kohaku-Blueleaf | 2023-10-25 | 1 | -1/+3 |
| | | | |||||
* | | | Fix alphas cumprod | Kohaku-Blueleaf | 2023-10-25 | 1 | -1/+2 |
| | | | |||||
* | | | Fix alphas_cumprod dtype | Kohaku-Blueleaf | 2023-10-25 | 1 | -0/+1 |
| | | | |||||
* | | | fp8 for TE | Kohaku-Blueleaf | 2023-10-25 | 1 | -0/+7 |
| | | | |||||
* | | | Fix lint | Kohaku-Blueleaf | 2023-10-23 | 1 | -1/+1 |
| | | | |||||
* | | | Add CPU fp8 support | Kohaku-Blueleaf | 2023-10-23 | 1 | -4/+16 |
| | | | | | | | | | | | | | | | | | | Since norm layer need fp32, I only convert the linear operation layer(conv2d/linear) And TE have some pytorch function not support bf16 amp in CPU. I add a condition to indicate if the autocast is for unet. | ||||
* | | | Add sdxl only arg | Kohaku-Blueleaf | 2023-10-19 | 1 | -0/+3 |
| | | | |||||
* | | | Add fp8 for sd unet | Kohaku-Blueleaf | 2023-10-19 | 1 | -0/+3 |
| |/ |/| | |||||
* | | repair unload sd checkpoint button | AUTOMATIC1111 | 2023-10-15 | 1 | -12/+1 |
| | | |||||
* | | use shallow copy for #13535 | AUTOMATIC1111 | 2023-10-14 | 1 | -2/+1 |
| | | |||||
* | | Merge pull request #13535 from chu8129/dev | AUTOMATIC1111 | 2023-10-14 | 1 | -4/+5 |
|\ \ | | | | | | | fix: checkpoints_loaded:{checkpoint:state_dict}, model.load_state_dict issue in dict value empty | ||||
| * | | reverst | wangqiuwen | 2023-10-07 | 1 | -0/+2 |
| | | | |||||
| * | | up | wangqiuwen | 2023-10-07 | 1 | -6/+5 |
| | | | |||||
* | | | Merge pull request #13139 from AUTOMATIC1111/ckpt-dir-path-separator | AUTOMATIC1111 | 2023-09-30 | 1 | -2/+3 |
|\ \ \ | | | | | | | | | fix `--ckpt-dir` path separator and option use `short name` for checkpoint dropdown | ||||
| * | | | parsing string to path | w-e-w | 2023-09-08 | 1 | -2/+3 |
| | | | | |||||
| * | | | keep order in list of checkpoints when loading model that doesn't have a ↵ | AUTOMATIC1111 | 2023-08-30 | 1 | -1/+21 |
| | | | | | | | | | | | | | | | | checksum | ||||
* | | | | add missing import, simplify code, use patches module for #13276 | AUTOMATIC1111 | 2023-09-30 | 1 | -7/+12 |
| | | | | |||||
* | | | | Merge pull request #13276 from woweenie/patch-1 | AUTOMATIC1111 | 2023-09-30 | 1 | -1/+14 |
|\ \ \ \ | |_|/ / |/| | | | patch DDPM.register_betas so that users can put given_betas in model yaml | ||||
| * | | | patch DDPM.register_betas so that users can put given_betas in model yaml | woweenie | 2023-09-15 | 1 | -1/+14 |
| | |/ | |/| | |||||
* | | | fix | 王秋文/qwwang | 2023-09-18 | 1 | -0/+1 |
| | | | |||||
* | | | use dict[key]=model; did not update orderdict order, should use move to end | qiuwen.wang | 2023-09-15 | 1 | -0/+1 |
|/ / | |||||
* / | keep order in list of checkpoints when loading model that doesn't have a ↵ | AUTOMATIC1111 | 2023-08-30 | 1 | -1/+21 |
|/ | | | | checksum | ||||
* | set devices.dtype_unet correctly | AUTOMATIC1111 | 2023-08-23 | 1 | -1/+2 |
| | |||||
* | add --medvram-sdxl | AUTOMATIC1111 | 2023-08-22 | 1 | -8/+8 |
| | |||||
* | Fix for consistency with shared.opts.sd_vae of UI | Uminosachi | 2023-08-21 | 1 | -0/+1 |
| | |||||
* | Change where VAE state are stored in model | Uminosachi | 2023-08-20 | 1 | -2/+0 |
| | |||||
* | Change to access sd_model attribute with dot | Uminosachi | 2023-08-20 | 1 | -2/+2 |
| | |||||
* | Store base_vae and loaded_vae_file in sd_model | Uminosachi | 2023-08-20 | 1 | -16/+8 |
| | |||||
* | Fix SD VAE switch error after model reuse | Uminosachi | 2023-08-20 | 1 | -2/+20 |
| | |||||
* | resolve the issue with loading fp16 checkpoints while using --no-half | AUTOMATIC1111 | 2023-08-17 | 1 | -1/+4 |
| | |||||
* | send weights to target device instead of CPU memory | AUTOMATIC1111 | 2023-08-16 | 1 | -1/+16 |
| | |||||
* | Revert "send weights to target device instead of CPU memory" | AUTOMATIC1111 | 2023-08-16 | 1 | -1/+1 |
| | | | | This reverts commit 0815c45bcdec0a2e5c60bdd5b33d95813d799c01. | ||||
* | send weights to target device instead of CPU memory | AUTOMATIC1111 | 2023-08-16 | 1 | -1/+1 |
| | |||||
* | put refiner into main UI, into the new accordions section | AUTOMATIC1111 | 2023-08-12 | 1 | -0/+3 |
| | | | | | | add VAE from main model into infotext, not from refiner model option to make scripts UI without gr.Group fix inconsistencies with refiner when usings samplers that do more denoising than steps | ||||
* | resolve merge issues | AUTOMATIC1111 | 2023-08-10 | 1 | -2/+5 |
| | |||||
* | Merge branch 'dev' into refiner | AUTOMATIC1111 | 2023-08-10 | 1 | -11/+12 |
|\ |