InvokeAI
InvokeAI copied to clipboard
invoke-ai
[bug]: ModuleNotFoundError: No module named 'ldm'
Is there an existing issue for this?
- [X] I have searched the existing issues
OS
Windows
GPU
cuda
VRAM
11
What happened?
After installing using Install.bat, running Invoke.bat adds this error.
Do you want to generate images using the
1. command-line
2. browser-based UI
3. open the developer console
Please enter 1, 2 or 3: 2
Starting the InvokeAI browser-based UI..
Traceback (most recent call last):
File "E:\invokeAI\scripts\invoke.py", line 2, in <module>
import ldm.invoke.CLI
ModuleNotFoundError: No module named 'ldm'
Adding sys.path.append(".") to invoke.py fixes this, but than the next module is not loaded. It seems the "scripts" folder, doesn't recognize all the modules from the root folder.
Screenshots
Additional context
No response
Contact Details
I was able to fix this by editing invoke.bat and adding this to the file (has to be done after install.bat as that file is overwritten during install)
set PYTHONPATH=%PYTHONPATH%;%cd%
However then I get a failure of (actual path replaced with <path>)
Do you want to generate images using the
1. command-line
2. browser-based UI
3. open the developer console
Please enter 1, 2 or 3: 2
Starting the InvokeAI browser-based UI..
Traceback (most recent call last):
File "C:\<path>\invokeAI\scripts\invoke.py", line 3, in <module>
import ldm.invoke.CLI
File "C:\<path>\invokeAI\ldm\invoke\CLI.py", line 9, in <module>
import yaml
ModuleNotFoundError: No module named 'yaml'
I think these are different issues but maybe related? Also on Windows
I had the same problem, and added set PYTHONPATH=%PYTHONPATH%;%cd% too, but instead of yaml error, this is what I got:
Starting the InvokeAI browser-based UI..
* Initializing, be patient...
>> Initialization file C:\Users\Barafu/.invokeai found. Loading...
>> InvokeAI runtime directory is "C:\Users\Barafu\invokeai"
Downloading: "https://github.com/DagnyT/hardnet/raw/master/pretrained/train_liberty_with_aug/checkpoint_liberty_with_aug.pth" to C:\Users\Barafu/.cache\torch\hub\checkpoints\checkpoint_liberty_with_aug.pth
100%|██████████████████████████████████████████████████████████████████████████████████████████████| 5.10M/5.10M [00:03<00:00, 1.64MB/s]
╭─────────────────────────────── Traceback (most recent call last) ────────────────────────────────╮
│ C:\Software\invokeAI\scripts\invoke.py:4 in <module> │
│ │
│ 1 #!/usr/bin/env python │
│ 2 │
│ 3 import ldm.invoke.CLI │
│ ❱ 4 ldm.invoke.CLI.main() │
│ 5 │
│ 6 │
│ │
│ C:\Software\invokeAI\ldm\invoke\CLI.py:54 in main │
│ │
│ 51 │ print(f'>> InvokeAI runtime directory is "{Globals.root}"') │
│ 52 │ │
│ 53 │ # loading here to avoid long delays on startup │
│ ❱ 54 │ from ldm.generate import Generate │
│ 55 │ │
│ 56 │ # these two lines prevent a horrible warning message from appearing │
│ 57 │ # when the frozen CLIP tokenizer is imported │
│ │
│ C:\Software\invokeAI\ldm\generate.py:133 in <module> │
│ │
│ 130 """ │
│ 131 │
│ 132 │
│ ❱ 133 class Generate: │
│ 134 │ """Generate class │
│ 135 │ Stores default values for multiple configuration items │
│ 136 │ """ │
│ │
│ C:\Software\invokeAI\ldm\generate.py:267 in Generate │
│ │
│ 264 │ │ ), 'call to img2img() must include the init_img argument' │
│ 265 │ │ return self.prompt2png(prompt, outdir, **kwargs) │
│ 266 │ │
│ ❱ 267 │ from ldm.invoke.generator.inpaint import infill_methods │
│ 268 │ │
│ 269 │ def prompt2image( │
│ 270 │ │ │ self, │
│ │
│ C:\Software\invokeAI\ldm\invoke\generator\inpaint.py:25 in <module> │
│ │
│ 22 infill_methods: list[str] = list() │
│ 23 │
│ 24 if Globals.try_patchmatch: │
│ ❱ 25 │ from patchmatch import patch_match │
│ 26 │ if patch_match.patchmatch_available: │
│ 27 │ │ print('>> Patchmatch initialized') │
│ 28 │ │ infill_methods.append('patchmatch') │
╰──────────────────────────────────────────────────────────────────────────────────────────────────╯
ModuleNotFoundError: No module named 'patchmatch'
I did a full reinstallation of everything - removed the invokeai folder, the anaconda environment... Everything except the downloaded models. Now it works without problems.
reinstall did not help me.. BUT I have probably 3 different stable diffusion installs so I wonder if the dependencies are conflicting in some way.
I did a clean installation of both, Python 3.10.9 (via the Microsoft Store) and InvokeAI on Windows 11/x64 (10.0.19045.0). I get the following error:
D:\invokeai> . .\invoke.bat
Do you want to generate images using the
1. command-line
2. browser-based UI
3. open the developer console
Please enter 1, 2 or 3: 2
Starting the InvokeAI browser-based UI..
Traceback (most recent call last):
File "D:\InvokeAI\invokeai\.venv\Scripts\invoke.py", line 3, in <module>
import ldm.invoke.CLI
ModuleNotFoundError: No module named 'ldm'
I get the same error using:
D:\InvokeAI\invokeai> & D:/InvokeAI/invokeai/.venv/Scripts/Activate.ps1
(.venv) D:\InvokeAI\invokeai\.venv> . ..\invoke.bat
However, when opening the invoke.py file with an IDE, the import is resolved without issues.
I guess there might be a problem with how the environment is set up during install.
I think the issue is with the Module Search Path. On my system this is:
D:\InvokeAI\invokeai> . .\invoke.bat
Do you want to generate images using the
1. command-line
2. browser-based UI
3. open the developer console
Please enter 1, 2 or 3: 3
Developer Console
Python command is:
C:\Users\movgp\AppData\Local\Microsoft\WindowsApps\python.exe
Python version is:
Python 3.10.9
*************************
You are now in the system shell, with the local InvokeAI Python virtual environment activated,
so that you can troubleshoot this InvokeAI installation as necessary.
*************************
*** Type `exit` to quit this shell and deactivate the Python virtual environment ***
(.venv) D:\InvokeAI\invokeai> python
>>> import sys
>>> sys.path
['',
'C:\\Program Files\\WindowsApps\\PythonSoftwareFoundation.Python.3.10_3.10.2544.0_x64__qbz5n2kfra8p0\\python310.zip',
'C:\\Program Files\\WindowsApps\\PythonSoftwareFoundation.Python.3.10_3.10.2544.0_x64__qbz5n2kfra8p0\\DLLs',
'C:\\Program Files\\WindowsApps\\PythonSoftwareFoundation.Python.3.10_3.10.2544.0_x64__qbz5n2kfra8p0\\lib',
'C:\\Users\\movgp\\AppData\\Local\\Microsoft\\WindowsApps\\PythonSoftwareFoundation.Python.3.10_qbz5n2kfra8p0',
'C:\\Users\\movgp\\AppData\\Local\\Packages\\PythonSoftwareFoundation.Python.3.10_qbz5n2kfra8p0\\LocalCache\\local-packages\\Python310\\site-packages',
'C:\\Users\\movgp\\AppData\\Local\\Packages\\PythonSoftwareFoundation.Python.3.10_qbz5n2kfra8p0\\LocalCache\\local-packages\\Python310\\site-packages\\win32',
'C:\\Users\\movgp\\AppData\\Local\\Packages\\PythonSoftwareFoundation.Python.3.10_qbz5n2kfra8p0\\LocalCache\\local-packages\\Python310\\site-packages\\win32\\lib',
'C:\\Users\\movgp\\AppData\\Local\\Packages\\PythonSoftwareFoundation.Python.3.10_qbz5n2kfra8p0\\LocalCache\\local-packages\\Python310\\site-packages\\Pythonwin',
'C:\\Program Files\\WindowsApps\\PythonSoftwareFoundation.Python.3.10_3.10.2544.0_x64__qbz5n2kfra8p0',
'C:\\Program Files\\WindowsApps\\PythonSoftwareFoundation.Python.3.10_3.10.2544.0_x64__qbz5n2kfra8p0\\lib\\site-packages'
]
However, the packages are located in
D:\InvokeAI\invokeai\.venv\Lib\site-packages
However, when opening the
invoke.pyfile with an IDE, the import is resolved without issues.
How to Resolve The Issue i'm Stuck With it not working via opening with IDE
Ok. Took me a while but finally found a solution.
I added an empty init.py (
) file to the scripts directory. This makes the scripts folder a part of a package and allows loading models from the other directories.
Ok. Took me a while but finally found a solution. I added an empty init.py (
👋 could you please help me I'm newbie stuck with that problem
Ok. Took me a while but finally found a solution. I added an empty init.py (
👋 could you please help me I'm newbie stuck with that problem
Simply go to the root folder where you installed InvokeAI, and you will see a folder there called "scripts". In it, create a file called "init.py". it should be an empty file. That should do the trick.
Not working for me
Ok. Took me a while but finally found a solution. I added an empty init.py (
👋 could you please help me I'm newbie stuck with that problem
Simply go to the root folder where you installed InvokeAI, and you will see a folder there called "scripts". In it, create a file called "init.py". it should be an empty file. That should do the trick.
Didn't work for me
Not working for me
Ok. Took me a while but finally found a solution. I added an empty init.py (
👋 could you please help me I'm newbie stuck with that problem
Simply go to the root folder where you installed InvokeAI, and you will see a folder there called "scripts". In it, create a file called "init.py". it should be an empty file. That should do the trick.
Didn't work for me
You made sure the file name includes underscores? also, please make sure your command line is with admin privileges. If both are true and you still can't make it work, please send a screenshot of the response you are getting.
Not working for me
Ok. Took me a while but finally found a solution. I added an empty init.py (
👋 could you please help me I'm newbie stuck with that problem
Simply go to the root folder where you installed InvokeAI, and you will see a folder there called "scripts". In it, create a file called "init.py". it should be an empty file. That should do the trick.
Didn't work for me
You made sure the file name includes underscores? also, please make sure your command line is with admin privileges. If both are true and you still can't make it work, please send a screenshot of the response you are getting.
After Step 4 When i Enter 1 or 2 it doest work but when i enter 3 it is working for developer console
When I installed I don't have a .venv folder. I have re-installed a few times and still no folder. I saw that a few of you all opened a file with an IDE or something. I assumed that kicked off some form of dependency update?
Do you any other steps than the normal install.bat command to create the .venv folder?
When I installed I don't have a .venv folder. I have re-installed a few times and still no folder. I saw that a few of you all opened a file with an IDE or something. I assumed that kicked off some form of dependency update?
Do you any other steps than the normal
install.batcommand to create the .venv folder?
I should Do Clean Installation
Not working for me
Ok. Took me a while but finally found a solution. I added an empty init.py (
👋 could you please help me I'm newbie stuck with that problem
Simply go to the root folder where you installed InvokeAI, and you will see a folder there called "scripts". In it, create a file called "init.py". it should be an empty file. That should do the trick.
Didn't work for me
You made sure the file name includes underscores? also, please make sure your command line is with admin privileges. If both are true and you still can't make it work, please send a screenshot of the response you are getting.
After Step 4 When i Enter 1 or 2 it doest work but when i enter 3 it is working for developer console
Don't put it in the venv script, rather in the root folder. this is the same folder where invoke.py resides:
Thanks @smoran I Found The Solution ! In The Begining i installed invokeai in a folder called "new" After Installation i renamed that folder
@smoran I've created a new file .\invokeai\.venv\Scripts\__init__.py, but running into the same issue. It's the only folder with an invoke.py file and there doesn't exist a .\invokeai\Scripts folder.
I ~~soved~~ workarounded the issue by making the following changes to .\invokeai\.venv\Scripts\invoke.py:
#!D:\invokeai\.venv\Scripts\python.exe
import sys
sys.path.insert(0,'D:\invokeai\.venv\Lib\site-packages')
import ldm.invoke.CLI
ldm.invoke.CLI.main()
I ran into the next issue with a concatenation of path names:
** model stable-diffusion-1.5 could not be loaded: [Errno 2] No such file or directory: 'D:\\InvokeAI\\invokeai\\models\\openai/clip-vit-large-patch14\\models--openai--clip-vit-large-patch14\\refs\\main'
Traceback (most recent call last):
File "D:\InvokeAI\invokeai\.venv\Lib\site-packages\ldm\invoke\model_cache.py", line 80, in get_model
requested_model, width, height, hash = self._load_model(model_name)
File "D:\InvokeAI\invokeai\.venv\Lib\site-packages\ldm\invoke\model_cache.py", line 233, in _load_model
model = instantiate_from_config(omega_config.model)
File "D:\InvokeAI\invokeai\.venv\Lib\site-packages\ldm\util.py", line 90, in instantiate_from_config
return get_obj_from_str(config['target'])(
File "D:\InvokeAI\invokeai\.venv\Lib\site-packages\ldm\models\diffusion\ddpm.py", line 658, in __init__
self.instantiate_cond_stage(cond_stage_config)
File "D:\InvokeAI\invokeai\.venv\Lib\site-packages\ldm\models\diffusion\ddpm.py", line 769, in instantiate_cond_stage
model = instantiate_from_config(config)
File "D:\InvokeAI\invokeai\.venv\Lib\site-packages\ldm\util.py", line 90, in instantiate_from_config
return get_obj_from_str(config['target'])(
File "D:\InvokeAI\invokeai\.venv\Lib\site-packages\ldm\modules\encoders\modules.py", line 247, in __init__
self.tokenizer = CLIPTokenizer.from_pretrained(
File "D:\InvokeAI\invokeai\.venv\Lib\site-packages\transformers\tokenization_utils_base.py", line 1760, in from_pretrained
resolved_vocab_files[file_id] = cached_file(
File "D:\InvokeAI\invokeai\.venv\Lib\site-packages\transformers\utils\hub.py", line 409, in cached_file
resolved_file = hf_hub_download(
File "D:\InvokeAI\invokeai\.venv\Lib\site-packages\huggingface_hub\utils\_validators.py", line 124, in _inner_fn
return fn(*args, **kwargs)
File "D:\InvokeAI\invokeai\.venv\Lib\site-packages\huggingface_hub\file_download.py", line 1150, in hf_hub_download
with open(ref_path) as f:
FileNotFoundError: [Errno 2] No such file or directory: 'D:\\InvokeAI\\invokeai\\models\\openai/clip-vit-large-patch14\\models--openai--clip-vit-large-patch14\\refs\\main'
The path does exist:
D:\InvokeAI\invokeai\models\openai\clip-vit-large-patch14\models--openai--clip-vit-large-patch14\refs> ls
Directory: D:\InvokeAI\invokeai\models\openai\clip-vit-large-patch14\models--openai--clip-vit-large-patch14\refs
Mode LastWriteTime Length Name
---- ------------- ------ ----
-a--- 14/12/2022 20:48 40 main
So I guess the / instead of the \\ is to blame here. Usually this comes from using string concatenation using + instead of os.path.join(...).
the easiest solution to this problem is to start the script and select the developers option then run pip install
I ran into this problem after updating today and it was missing the safetensors module...
Thanks @smoran I Found The Solution ! In The Begining i installed invokeai in a folder called "new" After Installation i renamed that folder
That was my problem as well. I had to update the VIRTUAL_ENV environment variable in activate.bat to reflect the new directory name.
CIRCUMSTANTIAL SOLUTION
I had this exact issue. I erroneously assumed that this was a stand alone app (in that it didn't have a file system "home" and could run from any arbitrary location) since it didn't have the usual windows install experience, when in fact, the directory it is installed in is hardcoded into many of the files during installation as an absolute path.
That said, I experienced the OP's issue because I had moved InvokeAI to a different location in my file system when I decided to consolidate my various SD UIs into one SD subfolder. Moving it back had it working properly. So my two choices, I believe, are ...
1. be happy with where it is
2. reinstall it at the location I really want it
Hope this helps others that in a similar situation.
