Comments (3)
You have to put the path were your checkpoint is.
if your model lays in a folder called "output" which resides in the same folder as train_dreambooth.py you would set it to MODEL_NAME="output".
accelerate config, depends on your setup.
How many CPUs and GPUs do you have?
Select that what fits you.
If you only got one GPU/CPU you can't do distributed training, and choose simply 0.
For deepspeed simply choose NO. Unless you know what you do.
For further help, read the documentation at https://huggingface.co/docs/accelerate/v0.12.0/en/package_reference/cli#accelerate-config
from diffusers.
Hey guys, I followed Nerdy Rodents video here and it solved all the issues. His video after that covers converting the output to a ckpt file also. I am on a 3080 10gb and 12 core 5900x and everything works great.
Make sure you use his pastebin link for the cuda install stuff and not the nvidia site links stuff because it's newer there and not compatible with bitsandbytes yet. It took me forever to realize what was happening. Now if I could only figure out how to use this all offline because when I disconnect internet it gives huggingface token https error during training. I tried disabling the token stuff in the .sh file but it doesn't allow training then. I really don't want to be forced to be connected when uploading pics of myself, call me paranoid.
One more tip was to make sure your training .sh file you upload is formatted for unix before sending it over to ubuntu, I had no idea this was a thing. In notepad++ you go to Edit, EOL Conversion, and choose Unix (LF) Before Windows was formatting the empty space with hidden characters and training wouldn't work. Hope this helps!
from diffusers.
Hey guys, I followed Nerdy Rodents video here and it solved all the issues. His video after that covers converting the output to a ckpt file also. I am on a 3080 10gb and 12 core 5900x and everything works great.
Make sure you use his pastebin link for the cuda install stuff and not the nvidia site links stuff because it's newer there and not compatible with bitsandbytes yet. It took me forever to realize what was happening. Now if I could only figure out how to use this all offline because when I disconnect internet it gives huggingface token https error during training. I tried disabling the token stuff in the .sh file but it doesn't allow training then. I really don't want to be forced to be connected when uploading pics of myself, call me paranoid.
One more tip was to make sure your training .sh file you upload is formatted for unix before sending it over to ubuntu, I had no idea this was a thing. In notepad++ you go to Edit, EOL Conversion, and choose Unix (LF) Before Windows was formatting the empty space with hidden characters and training wouldn't work. Hope this helps!
Can you help me i'm having some problems.
from diffusers.
Related Issues (20)
- RuntimeError: Detected that PyTorch and torchvision were compiled with different CUDA versions. HOT 8
- TypeError: Accelerator.__init__() got an unexpected keyword argument 'logging_dir' HOT 17
- AssertionError: You can't use same `Accelerator()` instance with multiple models when using DeepSpeed
- Why is the generated picture deformed? Why can't I generate a face picture that is the same as the original picture?
- COLAB BOG
- Colab training error
- Train multiple subjects in the same model HOT 3
- Setup for Paperspace.com
- Colab Fails to run half the time on a V100 HOT 1
- DreamBooths created with current version of Colab cannot be converted to LORAs in Kohya
- Please fix the notebook, it refuses to work. Install Requirements tab HOT 17
- Requirements error
- xformers wasn't built with CUDA support HOT 1
- Colab dreambooth notebook fail HOT 21
- Install Requirements (Fail)
- Install Requirements (Incompatible) and Exception: CUDA SETUP: Setup Failed!. HOT 2
- Training process fails with a Jax library related issue HOT 6
- torchvision problem nms does not exist HOT 2
- AttributeError: module 'jax.random' has no attribute 'KeyArray' while running DreamBooth_Stable_Diffusion.ipynb HOT 1
- Error training on SDXL 1.0 via Shivram's colab 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 diffusers.