| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
|
|
put queue stuff into separate file
|
|
|
|
completed, do use temporary dir to store it for the browser
|
|
|
|
resolve [name] after resolving [filewords] in training
|
|
Add CORS-allow policy launch argument using regex
|
|
Fix support for devices other than CUDA in DeepBooru
|
|
Fix api ignoring sampler_name settings
|
|
fix null negative_prompt on get requests
|
|
|
|
Forcing HTTPS instead of HTTP for ngrok
|
|
Clip skip + ENSD to parameters
|
|
|
|
|
|
Support NAI style exif in PNG Info for Send... buttons
|
|
Fix API img2img not accepting bare base64 strings
|
|
|
|
|
|
Supporting `*.safetensors` format.
|
|
Remove cmd args requirement for deepdanbooru
|
|
fix the model name error of Real-ESRGAN in the opts default value
|
|
add model_name pattern for saving
|
|
no-half support for SD 2.0
|
|
#3449 - VRAM leak when switching to/from inpainting checkpoint
|
|
Fix setting device for GFPGAN and CodeFormer
|
|
Fixes for PyTorch 1.12.1 when using MPS
|
|
|
|
|
|
|
|
torch.cuda.empty_cache() defaults to cuda:0 device unless explicitly …
|
|
|
|
|
|
add an option to choose a different temporary directory in the UI
add an option to cleanup the selected temporary directory at startup
|
|
otherwise first. Updating torch_gc() to use the device set by --device-id if specified to avoid OOM edge cases on multi-GPU systems.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Probably redundant, considering the model name *is* a filename, but I suppose better safe than sorry.
|
|
|
|
|
