CPU only fails because of missing NVIDIA driver

[VRciF]

Describe the bug

Right after txt2img finishes I receive the error

Total progress: 100%|██████████| 20/20 [07:53<00:00, 23.66s/it]
webui-docker-automatic1111-cpu-1  | Traceback (most recent call last):
webui-docker-automatic1111-cpu-1  |   File "/opt/conda/lib/python3.8/site-packages/gradio/routes.py", line 273, in run_predict
webui-docker-automatic1111-cpu-1  |     output = await app.blocks.process_api(
webui-docker-automatic1111-cpu-1  |   File "/opt/conda/lib/python3.8/site-packages/gradio/blocks.py", line 753, in process_api
webui-docker-automatic1111-cpu-1  |     result = await self.call_function(fn_index, inputs, iterator)
webui-docker-automatic1111-cpu-1  |   File "/opt/conda/lib/python3.8/site-packages/gradio/blocks.py", line 630, in call_function
webui-docker-automatic1111-cpu-1  |     prediction = await anyio.to_thread.run_sync(
webui-docker-automatic1111-cpu-1  |   File "/opt/conda/lib/python3.8/site-packages/anyio/to_thread.py", line 31, in run_sync
webui-docker-automatic1111-cpu-1  |     return await get_asynclib().run_sync_in_worker_thread(
webui-docker-automatic1111-cpu-1  |   File "/opt/conda/lib/python3.8/site-packages/anyio/_backends/_asyncio.py", line 937, in run_sync_in_worker_thread
webui-docker-automatic1111-cpu-1  |     return await future
webui-docker-automatic1111-cpu-1  |   File "/opt/conda/lib/python3.8/site-packages/anyio/_backends/_asyncio.py", line 867, in run
webui-docker-automatic1111-cpu-1  |     result = context.run(func, *args)
webui-docker-automatic1111-cpu-1  |   File "/stable-diffusion-webui/modules/ui.py", line 139, in f
webui-docker-automatic1111-cpu-1  |     mem_stats = {k: -(v//-(1024*1024)) for k,v in shared.mem_mon.stop().items()}
webui-docker-automatic1111-cpu-1  |   File "/stable-diffusion-webui/modules/memmon.py", line 77, in stop
webui-docker-automatic1111-cpu-1  |     return self.read()
webui-docker-automatic1111-cpu-1  |   File "/stable-diffusion-webui/modules/memmon.py", line 65, in read
webui-docker-automatic1111-cpu-1  |     free, total = torch.cuda.mem_get_info()
webui-docker-automatic1111-cpu-1  |   File "/opt/conda/lib/python3.8/site-packages/torch/cuda/memory.py", line 583, in mem_get_info
webui-docker-automatic1111-cpu-1  |     device = torch.cuda.current_device()
webui-docker-automatic1111-cpu-1  |   File "/opt/conda/lib/python3.8/site-packages/torch/cuda/__init__.py", line 481, in current_device
webui-docker-automatic1111-cpu-1  |     _lazy_init()
webui-docker-automatic1111-cpu-1  |   File "/opt/conda/lib/python3.8/site-packages/torch/cuda/__init__.py", line 216, in _lazy_init
webui-docker-automatic1111-cpu-1  |     torch._C._cuda_init()
webui-docker-automatic1111-cpu-1  | RuntimeError: Found no NVIDIA driver on your system. Please check that you have an NVIDIA GPU and installed a driver from http://www.nvidia.com/Download/index.aspx

Which UI

auto-cpu

Steps to Reproduce

1. From a fresh master git clone a few hours ago, I did docker compose --profile download up --build
2. docker compose --profile auto-cpu up --build
3. After this finishes, run the GUI at http://localhost:7860/ and without any modifications to settings use cat cartoon handrawing
4. A few images are generating during the processing which are shown in the WebUI, but at the very end (100%) I get above error.

Hardware / Software:

• OS: Ubuntu 22.04.1 LTS
• RAM: 16 GB
• GPU: NVIDIA Corporation GM107 [GeForce GTX 750 Ti]
• VRAM: 2GB (thus I'm using auto-cpu instead of gpu)
• Docker Version 20.10.18, Docker compose version 2.4.1
• Release version: master (6a66ff6)

Additional context

The WebUI only shows the message Error and in browser console I have Uncaught (in promise) API Error .
The final image is successfully saved to output directory.

@AbdBarho Many thanks for your efforts on this docker setup! You're doing an awesome job

[AbdBarho]

@VRciF Thanks for the detailed bug report! you have no idea how many people just completely ignore the template.

I have reached out to the maintainer behind the vram monitor feature, I have the impression they are a competent developer so this regression should be hopefully dealt with soon.

In the mean time, you can try older commits from master.

[AbdBarho]

@VRciF this should be fixed now, can you try again from latest master?

[VRciF]

I did git pull followed by docker compose --profile auto-cpu up --build resulting in a newly created container

 ⠿ Container webui-docker-automatic1111-cpu-1  Recreated                                                                                                                                                      0.1s
Attaching to webui-docker-automatic1111-cpu-1

and it is working well. I get no errors and the web ui works fine. Many, many thanks for your fast response!