This group is dedicated to the experimental AI Visualizer for Archicad 27, which will reach End of Service on December 31, 2024.
The new AI Visualizer in Archicad 28 is now fully cloud-based, with advanced controls like creativity and upscaling. Click here for further details.
2023-11-24 07:07 PM
Well, the AI Visualizer opens, and the code seems to be running fine on the terminal (at least following the user guide looks like it was supposed to be working)
Im using the Archicad 27 (Brazilian version if that matters) and a MacBook M2 Pro.
Any tips? Kinda new to archicad, been using it for a couple of months only
2023-11-24 08:58 PM
By the message at the end of the terminal windows looks like you run out of memory. Have you tried if reducing the Image Size has any effect?
2023-11-24 10:03 PM
@Aldo de Sa wrote:Im using the Archicad 27 (Brazilian version if that matters) and a MacBook M2 Pro.
Any tips?
Set the PYTORCH_MPS_HIGH_WATERMARK_RATIO environment variable to 0.4 or 0.6 and try again.
If you are unfamiliar with Terminal and environment variables, ask your more technical colleagues for help.
Example command in Terminal to be run in your sd.webui folder:
PYTORCH_MPS_HIGH_WATERMARK_RATIO=0.4 ./run.sh
or if you run runHighVram.sh:
PYTORCH_MPS_HIGH_WATERMARK_RATIO=0.4 ./runHighVram.sh
2023-11-27 11:14 AM
Hi Aldo de Sa,
Thanks for the question.
It looks like the process ran out of memory, please close all other apps running in the background and try to generate an image again.
Did you start the AI Engine in High Vram or in low Vram mode?
I suspect in HighVram, please try the add-on in low Vram.
On the AI Visulaizer UI:
- stop the engine (the button next to the 'Start AI Engine') and wait a minute
- uncheck the 'Start in High Vram mode' checkbox
- Start AI Engine (and wait for it to start)
- generate a picture
In case it still fails, please contact your local support and share the Terminal window output (Shell, Export text as...) as well.
Kind regards,
Technical Support Engineer
2023-11-30 11:25 AM
Got the following message in Terminal:
Stable diffusion model failed to load
Exception in thread Thread-2 (load_model):
Traceback (most recent call last):
File "/opt/homebrew/Cellar/python@3.10/3.10.13_1/Frameworks/Python.framework/Versions/3.10/lib/python3.10/threading.py", line 1016, in _bootstrap_inner
self.run()
File "/opt/homebrew/Cellar/python@3.10/3.10.13_1/Frameworks/Python.framework/Versions/3.10/lib/python3.10/threading.py", line 953, in run
self._target(*self._args, **self._kwargs)
File "/Applications/GRAPHISOFT/Archicad 27/sd.webui/webui/modules/initialize.py", line 153, in load_model
devices.first_time_calculation()
File "/Applications/GRAPHISOFT/Archicad 27/sd.webui/webui/modules/devices.py", line 148, in first_time_calculation
linear(x)
File "/opt/homebrew/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "/Applications/GRAPHISOFT/Archicad 27/sd.webui/webui/extensions-builtin/Lora/networks.py", line 429, in network_Linear_forward
return originals.Linear_forward(self, input)
File "/opt/homebrew/lib/python3.10/site-packages/torch/nn/modules/linear.py", line 114, in forward
return F.linear(input, self.weight, self.bias)
RuntimeError: "addmm_impl_cpu_" not implemented for 'Half'
2023-11-30 11:30 AM
I tried both settings with the same result.
Contacting the local support? I think this is a beta version and we are talking in a community,
Best regards,
Frank
2023-11-30 12:48 PM
Well, then go back to the clean start, but reduce the image size or the number of iterations in AI Visualizer add-on interface.
2023-11-30 12:49 PM
Hi Aixtreme,
Have you tried this?
On the AI Visulaizer UI:
- stop the engine (the button next to the 'Start AI Engine') and wait a minute
- uncheck the 'Start in High Vram mode' checkbox
- Start AI Engine (and wait for it to start)
- generate a picture
Please close all other apps running in the background to minimize memory usage.
In case it still fails, please share the Terminal window output (Shell, Export text as...) as well as a picture of the AI Add-on UI. Which macOS version are you on?
Thanks,
Technical Support Engineer
2023-11-30 01:44 PM
I will stop here with this Beta and wait for a final version, that is easier to install and working.
2024-01-03 12:21 PM - last edited on 2024-01-04 11:58 PM by Laszlo Nagy
I have the same problem.
I am running AC27 on an Apple M1 Max MacBook Pro with 32GB of RAM and Sonoma 14.2.1.
The AI visualiser, refuses to run even on the lowest settings and no other applications running.