Open camoody1 opened 1 year ago
Browser console output? I recently added support for tagger, but that logic should be optional.
I hope this is what you're asking for...
Could it have something to do with me using the RC version of Automatic1111? I have both the current release AND the release candidate version of A1111 installed on my machine. But I'm only running the RC version, right now.
Also, I have all of my models saved in a separate folder structure outside of the sdwebui folders since I also use ComfyUI and like having all of my downloaded files in one location.
Do you still have the flag on? The console log indicates that you are probably missing the cors flag.
set COMMANDLINE_ARGS=--api --cors-allow-origins https://huchenlei.github.io [Rest of ARGS...]
Yes. I still have that enabled in my webuser.bat file.
Any ideas? I'm using the release version of Automatic1111, again, and it's still not working. This is my webuser.bat file:
@echo off
set PYTHON= set GIT= set VENV_DIR= set COMMANDLINE_ARGS=--xformers --autolaunch --update-check --medvram --api --cors-allow-origins https://huchenlei.github.io ^ --ckpt-dir "H:\Stable_Diffusion_Models\models\stable-diffusion" ^ --vae-dir "H:\Stable_Diffusion_Models\models\VAE" ^ --gfpgan-dir "H:\Stable_Diffusion_Models\models\GFPGAN" ^ --esrgan-models-path "H:\Stable_Diffusion_Models\models\ESRGAN" ^ --bsrgan-models-path "H:\Stable_Diffusion_Models\models\BSRGAN" ^ --realesrgan-models-path "H:\Stable_Diffusion_Models\models\RealESRGAN" ^ --swinir-models-path "H:\Stable_Diffusion_Models\models\SwinIR" ^ --ldsr-models-path "H:\Stable_Diffusion_Models\models\LDSR" ^ --lora-dir "H:\Stable_Diffusion_Models\models\Lora" ^ --embeddings-dir "H:\Stable_Diffusion_Models\models\Embeddings" ^ --codeformer-models-path "H:\Stable_Diffusion_Models\models\Codeformer" ^ --controlnet-dir "H:\Stable_Diffusion_Models\models\ControlNet" git pull call webui.bat
I think it’s a1111’s issue. Other extensions using the api are also having problems. See https://github.com/AbdullahAlfaraj/Auto-Photoshop-StableDiffusion-Plugin/issues/361 . You can fall back to v1.5.2 for now while a1111 solving the issue.
I cannot reproduce this issue locally though.
hello have you tested it on Linux? i test it with 1.6 and 1.5.2 but no one connects, do you know if i have to make something extra? note: amazing work it gives so much power to photopea
I think the only thing you need to do is adding necessary commandline args when launching A1111.
--api --cors-allow-origins https://huchenlei.github.io
Do you have success with Windows but fail on Linux? Is the Linux a localhost or remote?
yes it was working perfect on windows with the directml version, i use this custom script for rocm to work
source venv/bin/activate export HSA_OVERRIDE_GFX_VERSION=10.3.0 export HIP_VISIBLE_DEVICES=0 export PYTORCH_HIP_ALLOC_CONF=garbage_collection_threshold:0.9,max_split_size_mb:512
python3 launch.py --api --cors-allow-origins https://huchenlei.github.io --enable-insecure-extension-access --opt-sdp-attention --no-half-vae --medvram
I'm running everything on local
Sorry I do not have a Linux machine to reproduce this issue, but you can verify that whether the cors-allow-origins is correctly configured by adding a print call here:
hello thx for the help i was testing and cors looks ok, but it still doesn't work very weird maybe some Linux behavior?, at the end i go with the public link option for a workaround.
"Connection Failed: TypeError: Failed to fetch"
I have Automatic1111 open with a live connection. I am able to create images in that application. However, when I try to connect in Photopea, I get that error message. Any suggestions to help resolve this?
I'm running the new Automatic1111 RC 1.6.0 if that makes a difference.