I'm confident that Dreambooth training will require less VRAM in the future, but currently the only lightweight implementations are Linux-only. Fixed prompt queue not working after running first entry Fixed an issue where the Stable Diffusion process would be killed when cancelling Model Pruning Tool: Added option to delete input file if pruning was successful Words in prompt field can now be deleted using CTRL+BACKSPACE, like in most text editors Added lots of hotkeys, documented on Github Readme Pop-Up Image Viewer: Added "Always on Top" option to keep window always on top Pop-Up Image Viewer: Window can now be resized/zoomed in 25% steps using mouse wheel Pop-Up Image Viewer: When slideshow mode is enabled, Left/Right arrow keys change images Added model folder manager: You can now add additional model folders to load models from Added Model Quick Switcher: Press Ctrl+M to quickly change the current AI model Added support for prompt emphasis - () for more, for less File "C:\Users\Me\AppData\Roaming\itch\apps\t2i-gui\SDG-1.9.- Added Dreambooth training (24 GB VRAM required for now!) return _load(opened_zipfile, map_location, pickle_module, **pickle_load_args) File "C:\Users\Me\AppData\Roaming\itch\apps\t2i-gui\SDG-1.9.1\Data\venv\lib\site-packages\torch\ serialization.py", line 712, in load return torch.load(checkpoint_file, map_location="cpu") File "C:\Users\Me\AppData\Roaming\itch\apps\t2i-gui\SDG-1.9.1\Data\venv\lib\site-packages\transformers\modeling_utils.py", line 399, in load_state_dict Canceled: Process has errored: Failed to load model. Implementation: InvokeAi - Force Kill: False Reason: Process has errored: Failed to load model. ** model sd-v1-5-fp16.ckpt-noVae could not be loaded: | Working with z of shape (1, 4, 32, 32) = 4096 dimensions. | Making attention of type 'vanilla' with 512 in_channels | LatentDiffusion: Running in eps-prediction mode > Loading sd-v1-5-fp16.ckpt-noVae from C:/Users/Me/AppData/Roaming/itch/apps/t2i-gui/SDG-1.9.1/Data/models/sd-v1-5-fp16.ckpt # NOT FOUND: CodeFormer model not found at. Loading Stable Diffusion with model "sd-v1-5-fp16". gfpgan -gfpgan_model_path gfpgan.pth -max_loaded_models 6 C:\Users\Me\AppData\Roaming\itch\apps\t2i-gui\SDG-1.9.1\Data\venv\Scripts\python.exe "repo/scripts/invoke.py" -o "C:/Users/Me/AppData/Roaming/itch/apps/t2i-gui/SDG-1.9.1/Data/sessions/-17-29-58/out" -precision float32 -gfpgan_dir. gfpgan -gfpgan_model_path gfpgan.pth -max_loaded_models 6 24290034 0' gfpgan -gfpgan_model_path gfpgan.pth -max_loaded_models 6 24290034 0' - New startup string: '-precision float32 -gfpgan_dir. Process running: False - Prev startup string: '-precision float32 -gfpgan_dir. 1 prompt * 1 image * 1 step value * 1 scale = 1 images total. InvokeAI model caching: Cache up to 5 models in RAM Running Stable Diffusion - 1 Iterations, 1 Steps, Scales 8, 512x512, Starting Seed: 1672315743 HasSdEnv - Has Python Exe: True - Has Pytorch: True HasBins - Has Python: True - Has Git: True - Has WKL: True - Has OK: True This window displays all messages that are logged while it's open. I've been experimenting with this GUI, which provides a user-friendly interface for stable diffusion, but recently encountered an error that says "Failed to load model." I'm not sure what's causing this issue, so I thought I'd reach out to the experts in this community for some guidance. I'm currently facing an issue with the NMKD Stable Diffusion GUI available on itch.io, and I could really use your help.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |