Closed kalle07 closed 3 months ago
i understand your frustration, but you're skipping over errors and making it worse by doing some steps manually.
You don't mention that you have to create a Venv first.
that is NOT needed for 99.9% users - all that webui.bat does is create and activate venv and then runs python launch.py
if that does not work on your system, that needs to be troubleshooted.
pip install -r requirements.txt
running requirements installation manually is NOT SUPPORTED. installer will install what it needs, you're just making it worse.
gpu dont work
webui.bat --use-cuda
will only install torch-cuda if torch is not already installed. in your case, you already had a bad installation.
so either remove venv or run with additional --reinstall
params.
Error running git: None / submodule --quiet update --init --recursive
this is the reason why k-diffusion is missing. details are in the log file which you did not provide.
stderr: 'fatal: detected dubious ownership in repository at 'E:/automatic/extensions/adetailer' 'E:/automatic/extensions/adetailer' is owned by: 'S-1-5-32-544' but the current user is: 'S-1-5-21-3859174244-1993617766-2322335723-1001'
make sure that you use same user to create folder, clone the repo and run webui.bat. do not switch to different users or use admin account.
no frustation , i dont understand why webui.bat dont work on my system ;)
and its heavy to troubleshoot if only ... c:\automatic>webui.bat c:\automatic> ... nothing happended ^^
ok i reinstalled it with activated venv webui.bat --reinstall
so now it works with cuda! :) (seems i bit faster than automatic1111)
i had to uncheck Full precision for model (--no-half) Full precision for VAE (--no-half-vae) 6GB SDXKL checkpoint load 16GB in VRAM (i have only 16)
Ensure that python and git are in your path by running python --version
and git --version
from the command line.
Then;
git clone https://github.com/vladmandic/automatic OptionalFolderName
cd automatic
(or cd OptionalFolderName
)
webui.bat
what do you mean "uncheck"? those settings are not enabled by default. they are typically only needed for gpus that don't support fp16. which gpu do you have?
and if webui.bat
just exits, remove line @echo off
at the start to see everything its doing
RTX4060 (for me it was checked maybe because my former instal and than reinstall)
e:\automatic2\automatic>webui.bat e:\automatic2\automatic>if not defined PYTHON (set PYTHON=python ) e:\automatic2\automatic>if not defined VENV_DIR (set "VENV_DIR=e:\automatic2\automatic\venv" ) e:\automatic2\automatic>set ERROR_REPORTING=FALSE e:\automatic2\automatic>mkdir tmp 2>NUL e:\automatic2\automatic>python -c "" 1>tmp/stdout.txt 2>tmp/stderr.txt e:\automatic2\automatic>python --version Python 3.10.5 e:\automatic2\automatic>git --version git version 2.40.1.windows.1
usualy all other python stuff work ;)
but it worked if i manual create an environment and activate it ? ;)
That's the thing, you shouldn't ever need to, since webui.bat does all that for you.
at least if webui makes all for me, an error should be occur (wrong python version) ^^
if this is last line that webui.bat prints, that means something terribly wrong happened to python, it did not even return error code
python -c "" 1>tmp/stdout.txt 2>tmp/stderr.txt
RTX4060 (for me it was checked maybe because my former instal and than reinstall)
i've asked several times to do clean install. there were way too many errors previously, especially with permissions. and why cuda was not detected is very questionable. not to mention you installing requirements manually.
all ok iam not a python programmer ;) thx for explaining
if i may ask, is all implemented if all is right installed and should download automaticaly? i try to go under tab control chose "Stable Video Diffusion" and error missing model iv dowloaded svd_xt_image_decoder.safetensors but which path?
(you can tell me if i should sart a new installation with python 3.10.11)
if i should sart a new installation with python 3.10.11
ideally, yes. in theory, 3.10.5 should also be fine but there is clearly something wrong there.
if i may ask, is all implemented if all is right installed and should download automaticaly?
only sd15 and sdxl are allowed to be downloaded manually as safetensors and be detected automatically. for other models, there is a lot of per-model configuration that needs to happen as safetensors does not contain sufficient data to even know what model type it is.
for svd, you should use extra networks -> reference -> select model and it will be downloaded automatically. and it can only be used in img2img with stable-video-diffusion script enabled.
i see...
btw my next try ...
e:\automatic>webui.bat --debug
e:\automatic>if not defined PYTHON (set PYTHON=python )
e:\automatic>if not defined VENV_DIR (set "VENV_DIR=e:\automatic\venv" )
e:\automatic>set ERROR_REPORTING=FALSE
e:\automatic>mkdir tmp 2>NUL
e:\automatic>python -c "" 1>tmp/stdout.txt 2>tmp/stderr.txt
e:\automatic>python --version Python 3.10.11
e:\automatic>git --version git version 2.40.1.windows.1
what can i do ? what is wrong and what is such as much different in automaitc1111 and fooocus and privatgpt and invoke and kohya and oobadooga(text-generation-webui) that i can install ? ;)
Did you try a full reinstall by following the steps here; https://github.com/vladmandic/automatic/issues/2972#issuecomment-1988676675
yes new git clone ...
Open command prompt and do where python
.
oh too many to choose
... C:\Users\kallemst>where python C:\Users\kallemst.pyenv\pyenv-win\shims\python C:\Users\kallemst.pyenv\pyenv-win\shims\python.bat C:\Users\kallemst\AppData\Local\Programs\Python\Python310\python.exe C:\Users\kallemst\AppData\Local\Microsoft\WindowsApps\python.exe
C:\Users\kallemst>
and? no idea ?
i would remove pyenv and python completely and install fresh python. my guess is that pyenv's python.bat is somehow preventing python check to complete
but only with your check ;)
all really all other work ... you have no manul for conda or environment install ? but nevertheless seems iam special case
but only with your check ;)
yes, sdnext performs many more checks than other similar apps since its the only one with true built-in support for different types of gpus out-of-the-box. that's why checks are necessary. if we were "its just cuda and for everything else you're on your own", then we could skip a lot of checks.
you have no manul for conda or environment install ?
if someone is using conda or pyenv i'd assume they have experience to set it up and troubleshoot. but yes, docs would be nice - and this is a community project, so contributions are welcome.
at the moment i studdy second time ^^
contributions is also share experience ;)
not sure what is expected here and now? i've asked to perform clean install and you've asked if you should use fresh python as well - i said yes. that was 3 days ago.
iv installed python 3.10.11 (old was .5) how you see last message
(you doo nothing with poetry ?)
i doo nothing more and risc all my other stuff ^^
what is the status of this issue? last i see was ask to see if it works with fresh python and without pyenv.
thx for asking,
i dont want delete all and reinstall pythion and git, i dont want risc my other 10 python repositories that working very well ... because im a beginner and i dont know what i doo ...
your check is to hard ;)
... E:\automatic>webui.bat
E:\automatic>if not defined PYTHON (set PYTHON=python )
E:\automatic>if not defined VENV_DIR (set "VENV_DIR=E:\automatic\venv" )
E:\automatic>set ERROR_REPORTING=FALSE
E:\automatic>mkdir tmp 2>NUL
E:\automatic>python -c "" 1>tmp/stdout.txt 2>tmp/stderr.txt
E:\automatic>
if i create venv python -m venv venv --system-site-packages
and activate it and run webui.bat than the install running ... it runs now, i can tell you if it worked ;)
i understand, but in that case not much i can do here and now.
it worked but only runs on CPU ... if its to compilate too set all manually ... you can close the issue
(venv) e:\automatic>webui
(venv) e:\automatic>if not defined PYTHON (set PYTHON=python )
(venv) e:\automatic>if not defined VENV_DIR (set "VENV_DIR=e:\automatic\venv" )
(venv) e:\automatic>set ERROR_REPORTING=FALSE
(venv) e:\automatic>mkdir tmp 2>NUL
(venv) e:\automatic>python -c "" 1>tmp/stdout.txt 2>tmp/stderr.txt
(venv) e:\automatic>if 0 == 0 goto :check_pip
(venv) e:\automatic>python -mpip --help 1>tmp/stdout.txt 2>tmp/stderr.txt
(venv) e:\automatic>if 0 == 0 goto :start_venv
(venv) e:\automatic>if ["e:\automatic\venv"] == ["-"] goto :skip_venv
(venv) e:\automatic>if [""] == ["1"] goto :skip_venv
(venv) e:\automatic>dir "e:\automatic\venv\Scripts\Python.exe" 1>tmp/stdout.txt 2>tmp/stderr.txt
(venv) e:\automatic>if 0 == 0 goto :activate_venv
(venv) e:\automatic>set PYTHON="e:\automatic\venv\Scripts\Python.exe"
(venv) e:\automatic>echo Using VENV: e:\automatic\venv
Using VENV: e:\automatic\venv
(venv) e:\automatic>if [] == ["True"] goto :accelerate
(venv) e:\automatic>goto :launch
(venv) e:\automatic>"e:\automatic\venv\Scripts\Python.exe" launch.py
16:34:07-177609 INFO Starting SD.Next
16:34:07-177609 INFO Logger: file="e:\automatic\sdnext.log" level=INFO size=65 mode=append
16:34:07-177609 INFO Python 3.10.11 on Windows
16:34:07-507680 INFO Version: app=sd.next updated=2024-03-21 hash=82973c49 branch=master
url=https://github.com/vladmandic/automatic/tree/master
16:34:08-037666 INFO Platform: arch=AMD64 cpu=Intel64 Family 6 Model 165 Stepping 5, GenuineIntel system=Windows
release=Windows-10-10.0.19045-SP0 python=3.10.11
16:34:08-048044 INFO nVidia CUDA toolkit detected: nvidia-smi present
16:34:08-048044 INFO Installing package: onnxruntime-gpu
16:34:37-327703 WARNING Modified files: ['webui.bat']
16:34:37-407658 INFO Startup: standard
16:34:37-407658 INFO Verifying requirements
16:34:37-417858 INFO Installing package: patch-ng
16:34:42-467859 INFO Installing package: addict
16:34:45-739121 INFO Installing package: astunparse
16:34:49-637999 INFO Installing package: blendmodes
16:35:00-430231 INFO Installing package: clean-fid
16:35:04-297841 INFO Installing package: filetype
16:35:08-127785 INFO Installing package: future
16:35:12-762721 INFO Installing package: httpcore
16:35:16-217698 INFO Installing package: inflection
16:35:19-937542 INFO Installing package: jsonmerge
16:35:26-027627 INFO Installing package: kornia
16:35:32-251911 INFO Installing package: lark
16:35:35-717804 INFO Installing package: lpips
16:35:39-527985 INFO Installing package: omegaconf
16:35:45-872586 INFO Installing package: open-clip-torch
16:35:52-947778 INFO Installing package: optimum
16:36:07-577807 INFO Installing package: piexif
16:36:11-540825 INFO Installing package: resize-right
16:36:15-407681 INFO Installing package: tensordict==0.1.2
16:36:20-392943 INFO Installing package: toml
16:36:24-317643 INFO Installing package: torchdiffeq
16:36:27-938547 INFO Installing package: voluptuous
16:36:31-457747 INFO Installing package: yapf
16:36:35-357952 INFO Installing package: scikit-image
16:36:45-517614 INFO Installing package: fasteners
16:36:49-167809 INFO Installing package: dctorch
16:36:52-957970 INFO Installing package: pymatting
16:36:57-227483 INFO Installing package: peft
16:37:02-157890 INFO Installing package: httpx==0.24.1
16:37:07-057642 INFO Installing package: compel==2.0.2
16:37:14-447439 INFO Installing package: torchsde==0.2.6
16:37:19-181085 INFO Installing package: clip-interrogator==0.6.0
16:37:23-627916 WARNING Package version mismatch: tqdm 4.66.2 required 4.66.1
16:37:23-637596 INFO Installing package: tqdm==4.66.1
16:37:27-477503 INFO Installing package: opencv-contrib-python-headless==4.9.0.80
16:37:38-547458 WARNING Package version mismatch: diffusers 0.27.2 required 0.27.0
16:37:38-547458 INFO Installing package: diffusers==0.27.0
16:37:45-207470 WARNING Package version mismatch: einops 0.7.0 required 0.4.1
16:37:45-207470 INFO Installing package: einops==0.4.1
16:37:49-377513 INFO Installing package: gradio==3.43.2
16:38:07-097677 INFO Installing package: numexpr==2.8.8
16:38:11-472517 WARNING Package version mismatch: numba 0.59.1 required 0.59.0
16:38:11-472517 INFO Installing package: numba==0.59.0
16:38:19-297923 WARNING Package version mismatch: protobuf 4.25.3 required 3.20.3
16:38:19-297923 INFO Installing package: protobuf==3.20.3
16:38:23-577531 INFO Installing package: pytorch_lightning==1.9.4
16:38:32-062335 INFO Installing package: tomesd==0.1.3
16:38:36-417873 WARNING Package version mismatch: urllib3 2.2.1 required 1.26.18
16:38:36-417873 INFO Installing package: urllib3==1.26.18
16:38:40-980538 WARNING Package version mismatch: timm 0.9.16 required 0.9.12
16:38:40-980538 INFO Installing package: timm==0.9.12
16:38:46-767751 WARNING Package version mismatch: pydantic 2.6.4 required 1.10.13
16:38:46-767751 INFO Installing package: pydantic==1.10.13
16:38:51-837855 WARNING Package version mismatch: typing-extensions 4.10.0 required 4.9.0
16:38:51-847887 INFO Installing package: typing-extensions==4.9.0
16:38:55-783227 INFO Verifying packages
16:38:55-783227 INFO Installing package: git+https://github.com/openai/CLIP.git
16:39:03-027655 INFO Installing package:
git+https://github.com/patrickvonplaten/invisible-watermark.git@remove_onnxruntime_depedency
16:39:20-317566 INFO Installing package: pi-heif
16:39:25-390292 INFO Installing package: tensorflow==2.13.0
16:40:52-446613 INFO Verifying submodules
16:43:35-487602 INFO Extension installed packages: sd-webui-controlnet ['dsine==2024.3.23',
'handrefinerportable==2024.2.12.0', 'depth-anything==2024.1.22.0', 'pywin32==306',
'svg.path==6.3', 'chardet==5.2.0', 'typing-extensions==4.10.0',
'opencv-contrib-python==4.9.0.80', 'opencv-python-headless==4.9.0.80', 'trimesh==4.2.4',
'reportlab==4.1.0', 'prettytable==3.10.0', 'embreex==2.17.7.post4', 'portalocker==2.8.2',
'shapely==2.0.3', 'numpy==1.26.4', 'onnx==1.16.0', 'easydict==1.13', 'svglib==1.5.1',
'albumentations==1.4.2', 'tinycss2==1.2.1', 'Rtree==1.2.0', 'geffnet==1.0.2',
'mapbox-earcut==1.0.1', 'mediapipe==0.10.11', 'tabulate==0.9.0', 'pycollada==0.8',
'insightface==0.7.3', 'cssselect2==0.7.0', 'webencodings==0.5.1', 'glob2==0.5', 'jax==0.4.25',
'sounddevice==0.4.6', 'ml-dtypes==0.4.0', 'iopath==0.1.9', 'yacs==0.1.8',
'fvcore==0.1.5.post20221221', 'xatlas==0.0.9', 'vhacdx==0.0.6']
16:43:40-537306 INFO Extension installed packages: stable-diffusion-webui-images-browser ['Send2Trash==1.8.2']
16:43:46-531091 INFO Extension installed packages: stable-diffusion-webui-rembg ['rembg==2.0.56']
16:43:46-533099 INFO Extensions enabled: ['Lora', 'sd-extension-chainner', 'sd-extension-system-info',
'sd-webui-agent-scheduler', 'sd-webui-controlnet', 'stable-diffusion-webui-images-browser',
'stable-diffusion-webui-rembg']
16:43:46-533099 INFO Verifying requirements
16:43:46-547269 WARNING Package version mismatch: typing-extensions 4.10.0 required 4.9.0
16:43:46-551302 INFO Installing package: typing-extensions==4.9.0
16:43:51-677591 INFO Command line args: []
16:44:03-557179 INFO Load packages: {'torch': '2.2.1+cpu', 'diffusers': '0.27.0', 'gradio': '3.43.2'}
16:44:06-367209 INFO Engine: backend=Backend.DIFFUSERS compute=cpu device=cpu attention="Scaled-Dot-Product"
mode=no_grad
16:44:06-367209 INFO Device:
16:44:08-407295 INFO Create: folder="models\ONNX"
16:44:08-417502 INFO Create: folder="models\Stable-diffusion"
16:44:08-417502 INFO Create: folder="models\Diffusers"
16:44:08-417502 INFO Create: folder="models\VAE"
16:44:08-417502 INFO Create: folder="models\Lora"
16:44:08-417502 INFO Create: folder="models\embeddings"
16:44:08-417502 INFO Create: folder="models\hypernetworks"
16:44:08-427207 INFO Create: folder="outputs\text"
16:44:08-427207 INFO Create: folder="outputs\image"
16:44:08-427207 INFO Create: folder="outputs\control"
16:44:08-427207 INFO Create: folder="outputs\extras"
16:44:08-433019 INFO Create: folder="outputs\init-images"
16:44:08-433019 INFO Create: folder="outputs\grids"
16:44:08-433019 INFO Create: folder="outputs\save"
16:44:08-437554 INFO Create: folder="outputs\video"
16:44:08-497248 INFO Available VAEs: path="models\VAE" items=0
16:44:08-507415 INFO Disabled extensions: ['sd-webui-controlnet']
16:44:08-507415 INFO Available models: path="models\Stable-diffusion" items=0 time=0.00
16:44:08-517255 INFO Installing package: basicsr
16:44:20-283575 INFO Installing package: gfpgan
16:44:29-327575 INFO Extension: script='extensions-builtin\Lora\scripts\lora_script.py'
16:44:29-327575 INFO LoRA networks: available=0
folders=2
16:44:30-247185 INFO Extension: script='extensions-builtin\sd-webui-agent-scheduler\scripts\task_scheduler.py'
Using sqlite file: extensions-builtin\sd-webui-agent-scheduler\task_scheduler.sqlite3
16:44:30-597159 INFO Extension:
script='extensions-builtin\stable-diffusion-webui-images-browser\scripts\image_browser.py'
Image Browser: Creating database
16:44:30-597159 INFO Extension:
script='extensions-builtin\stable-diffusion-webui-images-browser\scripts\image_browser.py'
Image Browser: Database created
16:44:30-682379 INFO Upscaler create: folder="models\chaiNNer"
16:44:30-682379 INFO Upscaler create: folder="models\RealESRGAN"
16:44:30-707200 INFO UI theme: name="black-teal" style=Auto base=sdnext.css
16:44:31-217639 ERROR Reading failed: ui-config.json [Errno 2] No such file or directory: 'ui-config.json'
16:44:32-034510 INFO Extension list is empty: refresh required
16:44:33-537458 INFO Local URL: http://127.0.0.1:7860/
16:44:33-727448 INFO [AgentScheduler] Task queue is empty
16:44:33-737526 INFO [AgentScheduler] Registering APIs
16:44:33-867463 WARNING Cannot generate without a checkpoint
16:44:33-867463 INFO Set system paths to use existing folders in a different location
16:44:33-867463 INFO or use --models-dir <path-to-folder> to specify base folder with all models
16:44:33-867463 INFO or use --ckpt-dir <path-to-folder> to specify folder with sd models
16:44:33-867463 INFO or use --ckpt <path-to-checkpoint> to force using existing model
16:44:33-877543 INFO Startup time: 42.19 torch=9.68 olive=0.11 gradio=2.08 libraries=4.80 ldm=0.05 samplers=0.09
extensions=1.41 face-restore=20.71 upscalers=0.06 ui-en=0.17 ui-txt2img=0.06 ui-img2img=0.07
ui-control=0.11 ui-settings=0.40 ui-extensions=1.49 ui-defaults=0.09 launch=0.37 api=0.10
app-started=0.22
16:51:19-406986 INFO Browser session: user=None client=127.0.0.1 agent=Mozilla/5.0 (Windows NT 10.0; Win64; x64;
rv:124.0) Gecko/20100101 Firefox/124.0
16:51:19-437019 INFO MOTD: N/A
16:53:57-607027 INFO Settings: changed=2 ['ckpt_dir', 'lora_dir']
16:54:03-856906 INFO Available models: path="d:\stable-diffusion\webui\models\Stable-diffusion\" items=102
time=0.72
16:54:29-496954 INFO Select: model="SDXL_myown_Sevenof9_VAE"
16:54:29-806985 WARNING Torch FP16 test failed: Forcing FP32 operations: "LayerNormKernelImpl" not implemented for
'Half'
16:54:29-816973 INFO Torch override dtype: no-half set
16:54:29-816973 INFO Torch override VAE dtype: no-half set
16:54:29-816973 INFO Setting Torch parameters: device=cpu dtype=torch.float32 vae=torch.float32 unet=torch.float32
context=no_grad fp16=False bf16=None optimization=Scaled-Dot-Product
16:54:29-822152 INFO Autodetect: model="Stable Diffusion XL" class=StableDiffusionXLPipeline
file="d:\stable-diffusion\webui\models\Stable-diffusion\SDXL_myown_Sevenof9_VAE.safetensors"
size=6617MB
Calculating hash: d:\stable-diffusion\webui\models\Stable-diffusion\SDXL_myown_Sevenof9_VAE.safetensors ---- 6.9… 0:00…
GB
16:54:47-547160 INFO Load embeddings: loaded=0 skipped=0 time=0.00
16:54:47-747017 INFO Load model: time=18.05 load=18.05 native=1024 {'ram': {'used': 13.6, 'total': 63.89}}
it worked but only runs on CPU ...
it runs on cpu because you have torch-cpu already installed. unless you specify --reinstall
, sdnext will not replace existing torch.
if its to compilate too set all manually
i dont understand?
i meant manually in the GUi
ok iv --reinstall(ed) webui --reinstall (with activated venv)
and ... IT works ;)
maybe one small issue for me if i input the resolution (with 4:3) i must doo it very slow, it calculate somewhat ?!? and if i change hight/width with that arrows, the resolution change not from 1024/768 to 768/1024 ?!?
i will try more ...
if i input the resolution (with 4:3) i must doo it very slow, it calculate somewhat ?!?
yes, its a known limitation - its a quite a slow ui update, will be improved in the future.
i see...
this resoltuion switch fron landscape to portrait you understand ? if not i can explain in new bug, and we can close here ;)
btw. i dont search for this in your gui yet extentions are aviable in a list (like automatic1111) ? and most are compatible ?
yes, they are available. compatible? depends.
Issue Description
You don't mention that you have to create a Venv first. than webui.bat runs ...
nevertheless after all is installed
11:01:27-423019 ERROR Error running git: None / submodule --quiet update --init --recursive 11:01:29-088543 ERROR Required path not found: path=E:\automatic\modules\k-diffusion\k_diffusion\sampling.py item=k_diffusion
see full log blow
Version Platform Description
win10 RTX4060
Relevant log output
Backend
Original
Branch
Master
Model
SD 1.5
Acknowledgements