New instructions for running web UI on macOS (check here for macOS related issues) #5461
Replies: 290 comments 771 replies
-
|
Hiyo, thanks so much for this! I'm happy to be a tester for this. I'm trying to get this setup on an M1 Max laptop; I removed a previous version that I'd installed with the "old" instructions (which didn't actually work; I had to do some file editing per this thread, which finally yielded a functional UI session). I removed all of that entirely and re-fetched the repo fresh following the above instructions. When I try to launch the UI now, I get: I'm not sure this is relevant, but: I use Python via pyenv, and currently have it set up to use 3.10.0 (not 3.10.6, as the 'main' instructions suggest, nor 3.10.8, which is what I have working on a separate Windows machine). Would that impact this? FWIW, prior to this, I also ran into two other issues trying to create TI embeddings, which I mention here. |
Beta Was this translation helpful? Give feedback.
-
|
I was getting this error:when I use hire.fix a image, the python begin to withdraw for no apparent reason and the console show a warning "/Library/Developer/CommandLineTools/Library/Frameworks/Python3.framework/Versions/3.9/lib/python3.9/multiprocessing/resource_tracker.py:216: UserWarning: resource_tracker: There appear to be 1 leaked semaphore objects to clean up at shutdown |
Beta Was this translation helpful? Give feedback.
-
|
I have no notes about the warning you're getting, but I've found in order for Refer to this comment to see my settings. Hope this helps! |
Beta Was this translation helpful? Give feedback.
-
|
I got this error too. And i change my model to a small one, and then it worked. |
Beta Was this translation helpful? Give feedback.
-
|
ModuleNotFoundError: No module named '_lzma' - FIXED Mac M1 Pro, Mac OS 14.1.2, Python (pyenv) 3.10.6 Couldn't get it to work following On first launch of ./webui.sh kept running into The one solution that made the application run was:
|
Beta Was this translation helpful? Give feedback.
-
|
Error = Error Domain=com.apple.appleneuralengine Code=6 "createProgramInstanceForModel:modelToken:qos:isPreCompiled:enablePowerSaving:skipPreparePhase:statsMask:memoryPoolID:enableLateLatch:modelIdentityStr:owningPid:cacheUrlIdentifier:aotCacheUrlIdentifier:error:: Program load failure (0x5)" UserInfo={NSLocalizedDescription=createProgramInstanceForModel:modelToken:qos:isPreCompiled:enablePowerSaving:skipPreparePhase:statsMask:memoryPoolID:enableLateLatch:modelIdentityStr:owningPid:cacheUrlIdentifier:aotCacheUrlIdentifier:error:: Program load failure (0x5)} What is the situation? After clicking to generate the image, there was no response... |
Beta Was this translation helpful? Give feedback.
-
|
I'd say including this is pretty important for Apple Silicon Macs. |
Beta Was this translation helpful? Give feedback.
-
|
Got some useful benchmarks about making one-off images too. For people running Stable Diffusion WebUI on battery, CoreML support is a must (600 -> 130 joules/image). Or even for the climate conscious, reducing emissions from 170 to 40 mg CO2 per image. |
Beta Was this translation helpful? Give feedback.
-
|
Guernika & Mochi Diffusion are good examples! I believe it is more easier to use already existing methods |
Beta Was this translation helpful? Give feedback.
-
|
The only downside is that mac os should be latest one - Ventura |
Beta Was this translation helpful? Give feedback.
-
|
Make no mistake, GPU rendering is faster than the neural engine. But it takes up more resources to do so. I've benchmarked both. |
Beta Was this translation helpful? Give feedback.
-
|
正克隆到 'stable-diffusion-webui'... |
Beta Was this translation helpful? Give feedback.
-
|
Much better instructions!
or something like that. I'm mentioning it only because I've read dozens of comments about people complaining about the space consumption issue. |
Beta Was this translation helpful? Give feedback.
-
|
i have this problem but still have no idea how to fix it |
Beta Was this translation helpful? Give feedback.
-
|
Copy or type in here the full path to your Auto1111 |
Beta Was this translation helpful? Give feedback.
-
|
@jrittvo I would really like to know how to do this also because my space on the main HD is getting very full! Can it be done in the startup ARGS? Here are my 2 links for the models:
Thanks my friend :) |
Beta Was this translation helpful? Give feedback.
-
|
First move or rename the /Users/briancurtis/stable-diffusion-webui/models/Stable-diffusion folder because it will be replaced by the symlink. Then, in Terminal: cd /Users/briancurtis/stable-diffusion-webui/models ln -s /Volumes/BEEJ2TB/Sept_Moved_SD_Models/Stable-diffusion |
Beta Was this translation helpful? Give feedback.
-
|
@jrittvo Thank You so much! :) This worked perfectly and now I can have ALL them accessible tgether and free-up . . . 200GB (cripes! I didn't realise I had tht many til I put them all tgether! lol) 😮🤪 Been looking for this answer for weeks now. 👍🏻⭐️⭐️⭐️ |
Beta Was this translation helpful? Give feedback.
-
|
Tried doing the "existing installation" instructions and got the following error after running webui.sh: Tried deleting the repositories and venv directories as instructred. Response from git pull was "already up to date" and it died in the same place with the same error (not that I expected success at that point). Completely trashed everything and tried a fresh install. That worked, though I'm still getting the |
Beta Was this translation helpful? Give feedback.
-
|
I'm getting the same error. Will try a fresh install as well. |
Beta Was this translation helpful? Give feedback.
-
|
I cloned the repo and followed the instructions; however, No module 'xformers'. Proceeding without it, still shows (even after I installed them using python -m pip install -U torch, python -m pip install -U xformers - the torch installation was needed as it could not find torch in my case). Were any of you able to resolve this issue? |
Beta Was this translation helpful? Give feedback.
-
|
Xformers doesn't run on macOS. |
Beta Was this translation helpful? Give feedback.
-
|
Thanks for the info |
Beta Was this translation helpful? Give feedback.
-
|
The BIG problem that I'm noticing is that, in a given session, performance degrades incredibly after the first batch. More details here: #5488 |
Beta Was this translation helpful? Give feedback.
-
|
I don't know if it helps, but I run a batch count of 100 every night with the following settings:
On my MBP M1 Pro with 32GB RAM, these settings in A1111 (last commit) generate a new image approx every 4min. Every morning, I check the terminal and every image is within the average. Except for one night, 2 days ago. Other than that, my generations are perfectly regular at approx 4min. |
Beta Was this translation helpful? Give feedback.
-
|
Before the 2.0 integration commit, I was getting that level of performance pretty consistently, and I AM STILL getting it pretty consistently in Invoke-AI. I'm consistently getting these incredibly long generation times in Auto for every generation passed the first in a given session. |
Beta Was this translation helpful? Give feedback.
-
|
I do notice that this appears to be launching into python 3.9.12, rather than the 3.10.x preferred by Invoke-AI. This is confirmed by observing memory use in Activity Monitor. This is especially odd given that i have confirmed via the brew info command that the Homebrew version of python installed is 3.10 as well. So questions:
|
Beta Was this translation helpful? Give feedback.
-
|
...except for still launching in py 3.9 unless I specify it in the command...i may just have to make a command file at this point. |
Beta Was this translation helpful? Give feedback.
-
|
so THIS is odd. When I run the following command everything launches fine: HOWEVER When I adjust webui-user.sh to introduce the exact same arguments in that command, it errors out with the following: |
Beta Was this translation helpful? Give feedback.
-
|
Seems that adjusting the python_cmd variable doesn't break things, so i've set that to 3.10. But enabling ANYTHING in the COMMANDLINE_ARGS variable causes the above assertion. |
Beta Was this translation helpful? Give feedback.
-
|
You’ll want to change that line to Also if you’ve updated web UI recently then it automatically sets |
Beta Was this translation helpful? Give feedback.
-
|
Huh...did a git pull earlier, so i should be up to date, but it's still defaulting to 3.9. No matter; easy enough to set in the webui_user file. |
Beta Was this translation helpful? Give feedback.
-
|
Does anyone have any insight into what would cause the errors in the terminal below after launching './webui.sh'? I'm running a 2021 MacBook Pro 16" (10c/24c w/ 64GB of RAM) on Ventura 13.0.1. Had one issue before this where I was unable to access "https://github.com/TencentARC/GFPGAN.git/", so I did it manually and then it worked the next time. Have tried clearing 'repositories' and 'venv' but still getting the same message. Edited for format |
Beta Was this translation helpful? Give feedback.
-
Actually no,
You might have a hidden extension on the file so it looks like the renamed file has the yaml extension when it still has an additional extension on it (e.g. Finder might show |
Beta Was this translation helpful? Give feedback.
-
|
You are completely right. I missed the |
Beta Was this translation helpful? Give feedback.
-
|
I had this issue, and while my configuration file only had one extension on it, I had downloaded it as a |
Beta Was this translation helpful? Give feedback.
-
|
I use someone else's |
Beta Was this translation helpful? Give feedback.
-
|
it seemed your model is not wor, you can only use ckpt or safesensors |
Beta Was this translation helpful? Give feedback.
-
|
Just followed this guide today and everything worked great. One note, I installed homebrew fresh, but I didn't have next steps for adding Homebrew to my PATH. Perhaps I had already done that previously with homebrew. |
Beta Was this translation helpful? Give feedback.
-
|
just put it in your .zshrc file |
Beta Was this translation helpful? Give feedback.
-
|
Tried training an embedding and it fails instantly. Get this error in the terminal: Traceback (most recent call last): |
Beta Was this translation helpful? Give feedback.
-
|
I hope someone would explain how it goes after launching the web and try training in hyper network and see how it goes. I don't see anyone who gets past that, there's no reference in YouTube or anywhere |
Beta Was this translation helpful? Give feedback.
-
|
See, I had just sort of assumed that this one of those bits that Don't Work On Mac. Though if this relies on xformers I would expect it to fail, as I'm reasonably sure the modular code needed hasn't been translated yet. |
Beta Was this translation helpful? Give feedback.
-
|
do you know of anyone who had this running completely functional. I'm past the installation, I could do text to image function but the main feature which is train images it failed and I'm stuck. so far I could only see people on YouTube running it on windows |
Beta Was this translation helpful? Give feedback.
-
|
In fact, I do not. |
Beta Was this translation helpful? Give feedback.
-
|
From my reply below:
Training is indeed broken at this time, but the PR should address that. |
Beta Was this translation helpful? Give feedback.
-
|
so I removed my old install, and started afresh. so I went over to hugging face, and tried to find one... now I get I got similar issues with the 2.1 model. |
Beta Was this translation helpful? Give feedback.
-
|
I’m assuming that’s the same as my issue? This reply from @brkirch fixed it:
|
Beta Was this translation helpful? Give feedback.
-
|
I am indeed. Thank you. |
Beta Was this translation helpful? Give feedback.
-
|
failed on step 5, after run ./webui.sh log as following: Using legacy 'setup.py install' for gfpgan, since package 'wheel' is not installed. stderr: Running command git clone --filter=blob:none --quiet https://github.com/TencentARC/GFPGAN.git /private/var/folders/d_/nz3s4q253752_l7h14mcc6cc0000gn/T/pip-req-build-8moh8ytq × Running setup.py install for lmdb did not run successfully. note: This error originates from a subprocess, and is likely not a problem with pip. × Encountered error while trying to install package. note: This is an issue with the package mentioned above, not pip. |
Beta Was this translation helpful? Give feedback.
-
|
It looks like there is a problem with your Xcode command line tools. Make sure Xcode is up to date in the App Store and then check Software Update in System Preferences or System Settings (whichever you have) to make sure you have all available software updates installed. |
Beta Was this translation helpful? Give feedback.
-
|
Hi, I reinstall my Xcode, and it works now. Thanks so much! |
Beta Was this translation helpful? Give feedback.
-
|
I also encountered such a problem, does this have anything to do with the system version? |
Beta Was this translation helpful? Give feedback.
-
|
I am experiencing the same issue and have not yet figured out how to resolve it. When you mention that the problem may be related to XCode, what exactly do you mean? I have not utilized XCode to install the web-ui and am not familiar with it. What should I do to solve the lmdb package error? Could someone please help me? (I am a newbie, so please be polite) |
Beta Was this translation helpful? Give feedback.
-
|
The real problem is "xcrun: error: unable to load libxcrun (dlopen(/Library/Developer/CommandLineTools/usr/lib/libxcrun.dylib, 0x0005): tried: ", this is related to Xcode tools, so I try the command "xcode-select --install", that works for me. |
Beta Was this translation helpful? Give feedback.
-
|
I am a completely non-technical person (a product manager) and I want to try out the tool to generate various images. I am wondering if anyone can provide instructions in layman's language with all (ALL, really) steps involved. I'd be more than happy to test, then create a document with screenshots and add various guides to the product. |
Beta Was this translation helpful? Give feedback.
-
|
If you are completely non-technical the easiest way to run Stable Diffusion would be to download DiffusionBee - https://joshdance.medium.com/how-to-run-stable-diffusion-on-an-m1-mac-a95185907289 - it is a standalone application, nothing to tweak, just enter commands in the text box and makes images. |
Beta Was this translation helpful? Give feedback.
-
|
i have chinese version of video about installation but for english version of guide no idea maybe you can find on on youtube teach you how to use |
Beta Was this translation helpful? Give feedback.
-
|
Maybe include the patch for k_diffusion to get more samplers working. |
Beta Was this translation helpful? Give feedback.
-
|
A k-diffusion fork based on Birch-san's k-diffusion fork is used for exactly that. Other than the PLMS sampler not working correctly with Stable Diffusion 2.0/2.1 models (this is also the case when running via CPU also so it may not be a Mac specific issue), all samplers should work correctly. |
Beta Was this translation helpful? Give feedback.
-
|
I'm using MacOS 12.6.1 Even after successful installation of gfpgan, it will be stuck on clip, then open_clip (if clip fortunately installs, probably 1 our of 5 times). |
Beta Was this translation helpful? Give feedback.
-
|
请问绕过防火墙的代码在安装sd的终端页面直接输入吗？我试了下还是失败了。 |
Beta Was this translation helpful? Give feedback.
-
yes, in china need vpn or clashxpro with enhance mode |
Beta Was this translation helpful? Give feedback.
-
Got it, thanks a lot! |
Beta Was this translation helpful? Give feedback.
-
我一开始按照层主的http.proxy方法也不行，后来重置代理后改为指定GitHub就可以了， |
Beta Was this translation helpful? Give feedback.
-
works! thanks a lot! |
Beta Was this translation helpful? Give feedback.
-
|
I get the following issue on my M1 as for some reason where venv installs x86 packages while python and arch is set to "ARM64". Any pointers on cleaning up or making the environment work on M1 for arm64? Somehow install the missing arm64 version of the crypto lib.
Launching Web UI with arguments: --no-half --use-cpu interrogate
Traceback (most recent call last):
File "/Users/miksvenson/repos/stable-diffusion-webui/launch.py", line 295, in <module>
start()
File "/Users/miksvenson/repos/stable-diffusion-webui/launch.py", line 286, in start
import webui
File "/Users/miksvenson/repos/stable-diffusion-webui/webui.py", line 11, in <module>
from modules.call_queue import wrap_queued_call, queue_lock, wrap_gradio_gpu_call
File "/Users/miksvenson/repos/stable-diffusion-webui/modules/call_queue.py", line 7, in <module>
from modules import shared
File "/Users/miksvenson/repos/stable-diffusion-webui/modules/shared.py", line 8, in <module>
import gradio as gr
File "/Users/miksvenson/repos/stable-diffusion-webui/venv/lib/python3.10/site-packages/gradio/__init__.py", line 3, in <module>
import gradio.components as components
File "/Users/miksvenson/repos/stable-diffusion-webui/venv/lib/python3.10/site-packages/gradio/components.py", line 31, in <module>
from gradio import media_data, processing_utils, utils
File "/Users/miksvenson/repos/stable-diffusion-webui/venv/lib/python3.10/site-packages/gradio/processing_utils.py", line 20, in <module>
from gradio import encryptor, utils
File "/Users/miksvenson/repos/stable-diffusion-webui/venv/lib/python3.10/site-packages/gradio/encryptor.py", line 2, in <module>
from Crypto.Cipher import AES
File "/Users/miksvenson/repos/stable-diffusion-webui/venv/lib/python3.10/site-packages/Crypto/Cipher/__init__.py", line 27, in <module>
from Crypto.Cipher._mode_ecb import _create_ecb_cipher
File "/Users/miksvenson/repos/stable-diffusion-webui/venv/lib/python3.10/site-packages/Crypto/Cipher/_mode_ecb.py", line 35, in <module>
raw_ecb_lib = load_pycryptodome_raw_lib("Crypto.Cipher._raw_ecb", """
File "/Users/miksvenson/repos/stable-diffusion-webui/venv/lib/python3.10/site-packages/Crypto/Util/_raw_api.py", line 309, in load_pycryptodome_raw_lib
raise OSError("Cannot load native module '%s': %s" % (name, ", ".join(attempts)))
OSError: Cannot load native module 'Crypto.Cipher._raw_ecb': Not found '_raw_ecb.cpython-310-darwin.so', Cannot load '_raw_ecb.abi3.so': cannot load library '/Users/miksvenson/repos/stable-diffusion-webui/venv/lib/python3.10/site-packages/Crypto/Util/../Cipher/_raw_ecb.abi3.so': dlopen(/Users/miksvenson/repos/stable-diffusion-webui/venv/lib/python3.10/site-packages/Crypto/Util/../Cipher/_raw_ecb.abi3.so, 0x0002): tried: '/Users/miksvenson/repos/stable-diffusion-webui/venv/lib/python3.10/site-packages/Crypto/Util/../Cipher/_raw_ecb.abi3.so' (mach-o file, but is an incompatible architecture (have 'x86_64', need 'arm64')), '/System/Volumes/Preboot/Cryptexes/OS/Users/miksvenson/repos/stable-diffusion-webui/venv/lib/python3.10/site-packages/Crypto/Util/../Cipher/_raw_ecb.abi3.so' (no such file), '/Users/miksvenson/repos/stable-diffusion-webui/venv/lib/python3.10/site-packages/Crypto/Util/../Cipher/_raw_ecb.abi3.so' (mach-o file, but is an incompatible architecture (have 'x86_64', need 'arm64')), '/Users/miksvenson/repos/stable-diffusion-webui/venv/lib/python3.10/site-packages/Crypto/Cipher/_raw_ecb.abi3.so' (mach-o file, but is an incompatible architecture (have 'x86_64', need 'arm64')), '/System/Volumes/Preboot/Cryptexes/OS/Users/miksvenson/repos/stable-diffusion-webui/venv/lib/python3.10/site-packages/Crypto/Cipher/_raw_ecb.abi3.so' (no such file), '/Users/miksvenson/repos/stable-diffusion-webui/venv/lib/python3.10/site-packages/Crypto/Cipher/_raw_ecb.abi3.so' (mach-o file, but is an incompatible architecture (have 'x86_64', need 'arm64')). Additionally, ctypes.util.find_library() did not manage to locate a library called '/Users/miksvenson/repos/stable-diffusion-webui/venv/lib/python3.10/site-packages/Crypto/Util/../Cipher/_raw_ecb.abi3.so', Not found '_raw_ecb.so' |
Beta Was this translation helpful? Give feedback.
-
|
Got the same error, but this didn't solve my error. Anybody? |
Beta Was this translation helpful? Give feedback.
-
|
It says it there; “is an incompatible architecture (have (x86_64), need (arm64e)))” at the bottom |
Beta Was this translation helpful? Give feedback.
-
|
Same issue. As suggested, clearing the cache and reinstalling didn't work for me, either. It seems like a gradio dependency (pycryptodome) is either:
As suggested in the issue, I configured Conda and ran: But no luck. Thank you world ahead of time for any help on this one! |
Beta Was this translation helpful? Give feedback.
-
|
To those with this issue: do you have Xcode installed? I suspect there is a problem with the build environment that is preventing pip from building the required wheel, but it is a bit hard to tell why exactly it isn't simply giving an error message when first trying to install pycryptodome instead of installing the wrong wheel. If it really is just pycryptodome that is the problem, then this wheel can be used: |
Beta Was this translation helpful? Give feedback.
-
|
Thanks a lot @brkirch, replacing the wheel manually as you proposed fixed the issue. The UI starts just fine now (MacBook Air M1, MacOS 13.1).
Yes, Xcode is installed. |
Beta Was this translation helpful? Give feedback.
-
SETTIGS内的这些标签无法点击，是什么问题？请大神赐教！
|
Beta Was this translation helpful? Give feedback.
-
|
I can't click on these tabs in the "settings", what is the problem? |
Beta Was this translation helpful? Give feedback.
-
|
Those are not tabs, but headers for the tabs below them. |
Beta Was this translation helpful? Give feedback.
-
|
Upgrading from 1.7 to 1.8 on M1 with Mac OS 13.6.4 throws this error at startup:
Seems to be a problem with Python3.9 vs. 3.10, see #15157 Is A1111 running in a virtual environment so that Python can be safely updated? |
Beta Was this translation helpful? Give feedback.
-
|
Download my a1111-setup script If your webui is in stable-diffusion-webui folder in your home directory, run it like this if it is some other folder, you can use -d switch to specify the folder location, for example After the webui is launched, upgrade all your plugins (important). The script will do all the steps from my step-by-step guide for you. It will remove your venv folder, upgrade your a1111 to latest version (using git command), apply some fixes and launch a1111. A1111 will do the rest and create a new venv folder for you. You can also follow my step-by-step guide. |
Beta Was this translation helpful? Give feedback.
-
|
Great Viking, I will check it! |
Beta Was this translation helpful? Give feedback.
-
|
Do you recommend the script over doing the pull yourself when updating? Tried to use it and now getting this at startup and can't track down where the imports are happening. After CL arguments ( " --skip-torch-cuda-test --upcast-sampling --no-half-vae --use-cpu interrogate --opt-split-attention --medvram-sdxl") comes ... launch.py 48 launch.py 44 main launch_utils.py 465 start webui.py 13 initialize.py 23 imports init.py 3 init.py 1 annotated_image.py 13 base.py 20 init.py 7 applications.py 15 routing.py 22 params.py 4 |
Beta Was this translation helpful? Give feedback.
-
|
这是来自QQ邮箱的假期自动回复邮件。
您好，您的邮件已收到。我会在考到消息后第一时间回复您。若内容紧要，请电话通知我，以免耽误！谢谢！
|
Beta Was this translation helpful? Give feedback.
-
|
macOS 13.6, 运行./webui.sh时报错 Python 3.10.13 (main, Jan 24 2024, 20:14:49) [Clang 15.0.0 (clang-1500.0.40.1)] Traceback (most recent call last): |
Beta Was this translation helpful? Give feedback.
-
You are having connection issues. Try to find and use a more stable connection. |
Beta Was this translation helpful? Give feedback.
-
|
I've had 1.6 running based on the instructions above on my M3 (MacOS 14.3.1, python 3.11.8), and I was able to update to 1.8 with just a One thing, however, isn't working for me: training embeddings. On my first try, it wouldn't load the blank(?) I ran it with that argument, it loaded, but when I attempted to train it, it gave the following output: For this run, my command-line arguments were:
Has anyone had better luck training a TI? |
Beta Was this translation helpful? Give feedback.
-
|
Python 3.10.x is recommended for A1111 for all OS. On macOS it is 3.10.13 |
Beta Was this translation helpful? Give feedback.
-
|
A1111 isn't the only app I have on this machine using Python. 3.10 receives no bug fixes, only security updates. 3.11 was released about a year and a half ago. Even 3.12 is almost 6 months old. Unless there's a specific reason the error I'm getting is due to 3.11 and there's a path forward for it, I'm not inclined to go through the pain of downgrading on that off chance. FWIW, it looks like support for 3.12 might be finally making its way here: |
Beta Was this translation helpful? Give feedback.
-
|
Your actual problem is this one: btw. have you also upgraded all your extensions as well after upgrading A1111? |
Beta Was this translation helpful? Give feedback.
-
|
hi everybody. tell me, I want to install facefusion in automatic1111, I go to extension - available - load form, and I see an error. can you tell me how to add face fusion without errors? |
Beta Was this translation helpful? Give feedback.
-
|
I met a strange problem ################################################################ and there's nothing continue, and after then show"time out service did not start " |
Beta Was this translation helpful? Give feedback.
-
|
Undata some messages ,it show the mistake. ################################################################ stderr: WARNING: Retrying (Retry(total=4, connect=None, read=None, redirect=None, status=None)) after connection broken by 'SSLError(SSLEOFError(8, 'EOF occurred in violation of protocol (_ssl.c:1007)'))': /simple/onnxruntime/ [notice] A new release of pip is available: 23.0.1 -> 24.0 Warning: Failed to install onnxruntime, some preprocessors may not work. stderr: WARNING: Retrying (Retry(total=4, connect=None, read=None, redirect=None, status=None)) after connection broken by 'SSLError(SSLEOFError(8, 'EOF occurred in violation of protocol (_ssl.c:1007)'))': /huchenlei/HandRefinerPortable/releases/download/v1.0.1/handrefinerportable-2024.2.12.0-py2.py3-none-any.whl [notice] A new release of pip is available: 23.0.1 -> 24.0 Warning: Failed to install handrefinerportable. Some processors will not work. stderr: WARNING: Retrying (Retry(total=4, connect=None, read=None, redirect=None, status=None)) after connection broken by 'SSLError(SSLEOFError(8, 'EOF occurred in violation of protocol (_ssl.c:1007)'))': /github-production-release-asset-2e65be/746748341/6542284e-6162-4441-b3aa-385ba9585b23?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=AKIAVCODYLSA53PQK4ZA%2F20240409%2Fus-east-1%2Fs3%2Faws4_request&X-Amz-Date=20240409T134505Z&X-Amz-Expires=300&X-Amz-Signature=90e44edb28172a51dbd8a041a2fd7dae05ce64b103e42f025a83b922e43cd7dc&X-Amz-SignedHeaders=host&actor_id=0&key_id=0&repo_id=746748341&response-content-disposition=attachment%3B%20filename%3Ddepth_anything-2024.1.22.0-py2.py3-none-any.whl&response-content-type=application%2Foctet-stream [notice] A new release of pip is available: 23.0.1 -> 24.0 Warning: Failed to install depth_anything. Some processors will not work. stderr: WARNING: Retrying (Retry(total=4, connect=None, read=None, redirect=None, status=None)) after connection broken by 'SSLError(SSLEOFError(8, 'EOF occurred in violation of protocol (_ssl.c:1007)'))': /sdbds/DSINE/releases/download/1.0.2/dsine-2024.3.23-py3-none-any.whl [notice] A new release of pip is available: 23.0.1 -> 24.0 Warning: Failed to install dsine. Some processors will not work. Installing requirements for TemporalKit extension Launching Web UI with arguments: --skip-torch-cuda-test --upcast-sampling --no-half-vae --use-cpu interrogate Civitai Helper: Root Path is: /Users/thalassocrat/stable-diffusion-webui To create a public link, set Startup time: 200.8s (launcher: 176.3s, import torch: 5.1s, import gradio: 1.5s, setup paths: 3.4s, other imports: 3.3s, setup codeformer: 0.1s, load scripts: 2.9s, create ui: 0.5s, gradio launch: 2.5s, app_started_callback: 5.1s). |
Beta Was this translation helpful? Give feedback.
-
|
@viking1304 Thank you so much for the installer, fantastic job. I was launching my last install with PYTORCH_MPS_HIGH_WATERMARK_RATIO=0.7 ./webui.sh --precision full --no-half. I could generate a SDXL image at 1024x1024 in just a few minutes. I did a fresh install of OS on my MacBook Pro M1 but now when I launch it with the same commend I get: `################################################################ ################################################################ ################################################################ I can launch it now with PYTORCH_MPS_HIGH_WATERMARK_RATIO=0.0 ./webui.sh --precision full --no-half, but my god it is PAINFULLY slow. Any suggestions? |
Beta Was this translation helpful? Give feedback.
-
|
Remove "--precision full --no-half" and do not use PYTORCH_MPS_HIGH_WATERMARK_RATIO. What is your exact Mac model? |
Beta Was this translation helpful? Give feedback.
-
|
Hi! |
Beta Was this translation helpful? Give feedback.
-
|
Try with my install script |
Beta Was this translation helpful? Give feedback.
-
|
When I attempt to run |
Beta Was this translation helpful? Give feedback.
-
|
try with those args in webui-user.sh |
Beta Was this translation helpful? Give feedback.
-
|
Hi, I also encountered the same situation. AttributeError: 'NoneType' object has no attribute 'lowvram' reinstallation also caused this situation. My version is Macos 12.7.4 Intel version, and your script seems to indicate that it is for situations related to 14.4.4 and the m series. What do I need to do if I want to modify it. Is the occurrence of AttributeError: 'NoneType' object has no attribute 'lowvram' a problem with my model? But I placed multiple models inside and the results were the same, |
Beta Was this translation helpful? Give feedback.
-
|
Still getting error messages |
Beta Was this translation helpful? Give feedback.
-
|
Hello, I installed stable diffusion on macos. It opens with webui, but when I create a prompt, Can I get a solution? ,,,Please.. *** Error completing request |
Beta Was this translation helpful? Give feedback.
-
|
Try using my install script. Let me know if it helps. |
Beta Was this translation helpful? Give feedback.
-
|
And a new problem... on my M3 MacBook Pro I can do inpainting (and acctually the important part, using adetailer) with SD 1.5 models without any problems at all but inpainting/adetailer with SDXL/Pony models does not work... I tried using --no-half which actually fix the problem, but then everything takes ages and that should not be the way to go (hmmm expect is there a way to activate --no-half just for the inpainting process... it would still be slow, but at least the normal image generation would still work in a speed as expected) Webui Version: v1.9.0 |
Beta Was this translation helpful? Give feedback.
-
|
Try with those args in webui-user.sh |
Beta Was this translation helpful? Give feedback.
-
|
That helped thanks, but I found another issue: This only works without a vae selected at all and whenever a vae is baked into the model (after i baked a vae into a model where I knew it would not have one), there is no luck getting it working with inpainting/adetailer at all... but still, being able to inpaint with sdxl/pony is a huge step forward ^^ |
Beta Was this translation helpful? Give feedback.
-
|
Apple M1Max, 32 GB, Sonoma 14.4.1. I have a problem with image generation. No error in the log. The problem manifests itself while using hires. fix. Specifically, at 60%, the image will completely destroyed and arrive at a result that is abolutely unrecognizable. To give you an idea, I'm recording a sped up video. zaznam1.mp4 |
Beta Was this translation helpful? Give feedback.
-
|
Try to uninstall it first then |
Beta Was this translation helpful? Give feedback.
-
|
I did right as you described. Uninstalling included. Just to be sure - I made it now again. Uninstall, then install. Still this: Python 3.10.14 (main, Mar 19 2024, 21:46:16) [Clang 15.0.0 (clang-1500.3.9.4)] |
Beta Was this translation helpful? Give feedback.
-
|
Sorry, I forgot about venv: |
Beta Was this translation helpful? Give feedback.
-
|
I found out I had to delete the venv first, using --reinstall-torch wasnt enough , kept on 2.3dev , I deleted the venv and that did it. Downloaded and installed 2.4 dev. In webui-user.sh I used (per viking recomendation) to create a venv for development torch to could fall back easily in case of problems: venv_dir="venv-torch-nightly" and: export TORCH_COMMAND="pip install --pre torch torchvision --index-url https://download.pytorch.org/whl/nightly/cpu" comment (#) all others related to torch. |
Beta Was this translation helpful? Give feedback.
-
|
viking1304: You are a lifesaver. After deleting venv, everything works like a charm! Thank you! |
Beta Was this translation helpful? Give feedback.
-
|
macbook pro: 2.8 GHz 四核Intel Core i7, Intel Iris Plus Graphics 655 1536 MB, 16 GB 2133 MHz LPDDR3 , txt to img is Very, very slow，Do you have this problem? |
Beta Was this translation helpful? Give feedback.
and others
-
Mac users: Please provide feedback on if these instructions do or don't work for you, and if anything is unclear or you are otherwise still having problems with your install that are not currently mentioned here.
Important notes
Currently most functionality in the web UI works correctly on macOS, with the most notable exceptions being CLIP interrogator and training. Although training does seem to work, it is incredibly slow and consumes an excessive amount of memory. CLIP interrogator can be used but it doesn't work correctly with the GPU acceleration macOS uses so the default configuration will run it entirely via CPU (which is slow).
Most samplers are known to work with the only exception being the PLMS sampler when using the Stable Diffusion 2.0 model. Generated images with GPU acceleration on macOS should usually match or almost match generated images on CPU with the same settings and seed.
Automatic installation
New install:
brew install cmake protobuf rust [email protected] git wgetgit clone https://github.com/AUTOMATIC1111/stable-diffusion-webuistable-diffusion-webui/models/Stable-diffusion. If you don't have any, see Downloading Stable Diffusion Models below.cd stable-diffusion-webuiand then./webui.shto run the web UI. A Python virtual environment will be created and activated using venv and any remaining missing dependencies will be automatically downloaded and installed../webui.shagain. Note that it doesn't auto update the web UI; to update, rungit pullbefore running./webui.sh.Existing Install:
If you have an existing install of web UI that was created with
setup_mac.sh, delete therun_webui_mac.shfile andrepositoriesfolder from yourstable-diffusion-webuifolder. Then rungit pullto update web UI and then./webui.shto run it.Downloading Stable Diffusion Models
If you don't have any models to use, Stable Diffusion models can be downloaded from Hugging Face. To download, click on a model and then click on the
Files and versionsheader. Look for files listed with the ".ckpt" or ".safetensors" extensions, and then click the down arrow to the right of the file size to download them.Some popular official Stable Diffusion models are:
Stable Diffusion 2.0 and 2.1 require both a model and a configuration file, and image width & height will need to be set to 768 or higher when generating images:
For the configuration file, hold down the option key on the keyboard and click here to download
v2-inference-v.yaml(it may download asv2-inference-v.yaml.yml). In Finder select that file then go to the menu and selectFile>Get Info. In the window that appears select the filename and change it to the filename of the model, except with the file extension.yamlinstead of.ckpt, press return on the keyboard (confirm changing the file extension if prompted), and place it in the same folder as the model (e.g. if you downloaded the768-v-ema.ckptmodel, rename it to768-v-ema.yamland put it instable-diffusion-webui/models/Stable-diffusionalong with the model).Also available is a Stable Diffusion 2.0 depth model (512-depth-ema.ckpt). Download the
v2-midas-inference.yamlconfiguration file by holding down option on the keyboard and clicking here, then rename it with the.yamlextension in the same way as mentioned above and put it instable-diffusion-webui/models/Stable-diffusionalong with the model. Note that this model works at image dimensions of 512 width/height or higher instead of 768.Troubleshooting
Web UI Won't Start:
If you encounter errors when trying to start the Web UI with
./webui.sh, try deleting therepositoriesandvenvfolders from yourstable-diffusion-webuifolder and then update web UI withgit pullbefore running./webui.shagain.Poor Performance:
Currently GPU acceleration on macOS uses a lot of memory. If performance is poor (if it takes more than a minute to generate a 512x512 image with 20 steps with any sampler) first try starting with the
--opt-split-attention-v1command line option (i.e../webui.sh --opt-split-attention-v1) and see if that helps. If that doesn't make much difference, then open the Activity Monitor application located in /Applications/Utilities and check the memory pressure graph under the Memory tab. If memory pressure is being displayed in red when an image is generated, close the web UI process and then add the--medvramcommand line option (i.e../webui.sh --opt-split-attention-v1 --medvram). If performance is still poor and memory pressure still red with that option, then instead try--lowvram(i.e../webui.sh --opt-split-attention-v1 --lowvram). If it still takes more than a few minutes to generate a 512x512 image with 20 steps with with any sampler, then you may need to turn off GPU acceleration. Openwebui-user.shin Xcode and change#export COMMANDLINE_ARGS=""toexport COMMANDLINE_ARGS="--skip-torch-cuda-test --no-half --use-cpu all".Beta Was this translation helpful? Give feedback.
All reactions