Comments (8)
it seems ran out of system ram rather than vram - probably we need a bit optimization on this ..
from fooocus.
Machine has 16GB of ram. I added 4gb of swap and it is now running! I'm a little concerned that the memory was never released after the install though.
from fooocus.
Sadly it is not usable. I do get the UI up and can enter a prompt but it starts to process then everything seems to freeze. Also, I see no movement on the GPU usage chart.
from fooocus.
yes we may confirm that this is low ram problem.
this is because both xl base and refiner are used for best results.
if invokeai can work on this machine, then does invokeai use both base and refiner?
if they are not loading less models then i can probably take a look at other methods for system ram optimization
from fooocus.
You're quite right. SDXL under InvokeAI doesn't run either. I will get a bigger instance tomorrow and go through it again.
from fooocus.
The test passed on my end with 8GB RAM on windows. see also updated readme installation guides with images and explanations.
But I am not sure if linux can do that
from fooocus.
I'm using a Proxmox cluster to try and run Fooocus on a Ubuntu VM. I pass through a RTX3060 with 12GB to that VM. The host machine has 192GB of RAM, with 64GB allocated to the VM.
This VM runs SDXL on InvokeAI, Auto111, ComfyUI, without issue.
When running python launch.py --listen
, the process is eventually killed. I can paste all the messages seen after executing the launch script, but ultimately there's no error - it just ends with:
building MemoryEfficientAttnBlock with 512 in_channels...
Killed
I'm not quite sure if this is a RAM issue or what. Typically when I've seen Killed
it's because of too little ram. That said, not sure why other SDXL compatible projects load the SDXL models without issue.
from fooocus.
I'm also getting:
building MemoryEfficientAttnBlock with 512 in_channels... [1] 3299 killed python launch.py
OS: Arch Linux
CPU: i5-9600k
GPU: GTX 750 ti
RAM; 24GB DDR4
Swap: 4GB
from fooocus.
Related Issues (20)
- Loras not working: ValueError: Error while deserializing header: MetadataIncompleteBuffer HOT 1
- [Bug]: custom preset fails HOT 2
- [Feature Request]: Using Florence 2 in describe option HOT 1
- [Bug]: No errors, not starting at all HOT 1
- [Feature Request]: Ability to pause/resume generations HOT 6
- [Feature Request]: Suggestions for adjusting log.html HOT 2
- [Bug]: Foocus not starting "ModuleNotFoundError: No module named 'supervision'" HOT 4
- [Bug]: API do not create Images HOT 1
- [Feature Request]: Add Support for Flux models HOT 1
- [Bug]: Plz fix A1111 Metadata scheme (with styles and Enhance) HOT 2
- [Bug]: Inpaint: Improve Detial - Not always work HOT 4
- [Feature Request]: Please add support for FLUX.1-dev HOT 1
- [Feature Request]: Plz add an option to export metadata for A1111/Forge/CivitAI HOT 1
- [Bug]: Inpaint or Outpaint->Improve Detail, Don´t work with the new update HOT 2
- Would width and height inputs be possible? [Feature Request]: HOT 1
- VRAM load doesn`t get lower [Bug]: HOT 1
- [Bug]: fooocus raising exception when load kolors checkpoint model HOT 1
- [Feature Request]: Add a "balanced" option in the performance section to sit between quality and speed HOT 3
- [Feature Request]: FaceSwap optimizer HOT 2
- [Feature Request]: LoRA notification support in prompt array HOT 1
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from fooocus.