🐛〉bugs-and-support
outcrop jpg image - Initial image has transparent areas. Will inpaint in these regions.
--fix on an image without metadata causes a crash
Problem when running custom embeddings via Windows Power Shell
img2img - Expected {conjunction | implicit_conjunction}, found '('
img2img - poor quality output
Missing icons compared to the 2.1 update video
Using -I with a PNG always gives an essentially identical image
Tool or command for extracting single images from grid image, retaining per-image EXIF data?
Huggingface token issue
Attempting to upgrade to 2.1 on M1 Mac [SOLVED]
image to image on v2.1
Can't load model after update
Getting failure with inpainting on M1 (MPS assert)
Documentation issues
Can't install on M1 Pro with conda, solving environment fails
Can't train Textual Inversion
Can't create conda env from environment.yml
Maybe a dumb question about switching models
Switching from CPU to GPU
Much slower than it used to be, and not using much GPU?
Corruption or Glitches
M1 Mac - GUI import launch error
Stuck installing pip dependencies when updating
invoke.py only works with sudo
I have a question on 64 * 64 resolution with 50 steps taking 10 minutes on M1 8GB.
having a problem with hash checks, would just like to disable them outright
Upscaling is creating dark images
Corrupted gfpgan file
Runtime Error after running (python scriptspreload_models.py)
Parentheses in prompts causing errors in latest from development branch
Using Device Type CPU
--embiggen can cause InvokeAI to crash
--seamless is not stored inside of an image's Sd-metadata but is present in the Dream metadata.
Having difficulty switching CPU to GPU on a macOS A1
Can't run the InvokeAI... install all active all..
Server error 'NoneType' object has no attribute 'modules'
hardware requirements in github repo says 4gb vram as minimum but just loading the model takes 4.56g
High Resolution Optimization (hires_fix) and initial dimensions less than 512
The dataloader, train_dataloader, does not have many workers which may be a bottleneck
Batches have the same seed via web api
Is it possible to save output during steps?
^C on CLI with `!fix [...] --embiggen [...]` terminates the script
Is it possible to use custom arguments with the web based gui?
Trying the understand the inconsistency wrt DreamStudio
How many epochs does a stable diffusion model train on textual inversion?
InvokeAI CodeFormer producing bad results
ModuleNotFoundError No module named 'models.clipseg' on Windows
Textual inversion seems broken
***FIXED*** Took over an Hour to generate 10 images and no Webui is this normal!??
What can I do when it says GFPGAN isn't found?