A diffusers based implementation of HyperDreamBooth
License: Apache License 2.0
🔭 I’m currently working on LyCORIS
🤝 I’m looking for help with HyperKohaku
💬 Ask me about Python, NN, Web Crawler
📫 How to reach me [email protected]
⚡ Fun fact I never watched Lycoris-Recoil
paypal.me: https://www.paypal.com/paypalme/kblueleaf
BTC: 36VHoCKxgp2u3YWQ8gNMDQR3fT49S5sRtf
ETH: 0x8023c8c0a10a4da4e6746cbd238a8bc990fbba60
LTC: MCpMKubB8eeKPZ6LsfW9A7pJP23YLoLT9T
I noticed that the original paper stated that down_aux/up_aux is a progressive orthogonal vector, but the code in this repository orthogonalized the auxiliary layer weights with down_dim/out_feature_dim. This may be the reason why losses cannot be converged. Can you take a look at it? Thanks!
Thank you for your work. What are the detailed steps for the fourth step?
What is the diffusers version, please! Thanks!
accelerate is configured with multiple gpus, but only one gpu is used for training
Hello, where do I find the celeba-hq-512-4.8k dataset?
I followed all the steps:
train_preoptimized_liloras.pytrain_hyperdreambooth.pyhypernetwork_gen_weigth.pyinference_test.py, but I get this error:Traceback (most recent call last):
File "/home/wizard/bendai/research/dawnai/hyper_dreambooth/inference_test.py", line 19, in <module>
pipe.unet.load_attn_procs(model_path)
File "/home/wizard/mambaforge/envs/hyper/lib/python3.9/site-packages/diffusers/loaders.py", line 297, in load_attn_procs
state_dict = safetensors.torch.load_file(model_file, device="cpu")
File "/home/wizard/mambaforge/envs/hyper/lib/python3.9/site-packages/safetensors/torch.py", line 259, in load_file
with safe_open(filename, framework="pt", device=device) as f:
safetensors_rust.SafetensorError: Error while deserializing header: HeaderTooLarge
Any idea? Thank you!
The loss function converges normally during Pre-Optimize training, but during the subsequent HyperNet training, both the gen_loss and weight_loss do not converge.
after exporting the HyperNet predicted weights to standard LoRA, the inference results are abnormal.
Hello, is the project complete and do you plan to add command examples for us to run the codes?
Thanks.
first thank you for progress awesome project !
I want to know about milestone or progress of this project
did you train hyper network model now? or
how many time take to train hyper network model?
Finally, based on the current state, I wonder if the current project is actually usable.
Could you provide information on the Torch versions that are compatible with this code?
Thx for your project!~
I find this sentence in readme: 'Implement better dreambooth training (like masking the face during training)'.
But how to mask the face during training? Eg: random mask 50%?
Would there be pretrained hypernetwork released ?
Thank you for this fantastic project.
I have some questions regarding the operational condition. What is the required GPU memory for this project? Does this project support multi-GPU training? How long does it take to achieve reasonably good results with only four 32GB V100?
I can't do anything
I have written a script to export the pre-optimized weights(pre_optimized.bin ) to LoRA according to designated ID , and tested it using inference_test.py, I found that the generated result is distorted. Can I provide my script so that everyone can verify it as well?
HyperKohaku/train_hyperdreambooth.py
Line 1050 in 7651313
I cant' find where weight_preds are loaded into T2I model for the reconstruction error
Hi! Thank you for the amazing work. Training on a single GPU is quite slow, and since the project uses accelerate, I was expecting it to run also on multiple GPUs. However, after some small tweaks (removing the --put_in_cpu flag when training the preoptimized loras, and substituting hypernetwork.train_params() with hypernetwork.module.train_params()) I am stuck with this error:
Traceback (most recent call last):
File "/home/hyper_dreambooth/./train_hyperdreambooth.py", line 1323, in <module>
main(args)
File "/home/hyper_dreambooth/./train_hyperdreambooth.py", line 1117, in main
pred_weights = hypernetwork(pixel_values)
File "/home/mambaforge/envs/hyper/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1501, in _call_impl
return forward_call(*args, **kwargs)
File "/home/mambaforge/envs/hyper/lib/python3.9/site-packages/torch/nn/parallel/distributed.py", line 1139, in forward
if torch.is_grad_enabled() and self.reducer._rebuild_buckets():
RuntimeError: Expected to have finished reduction in the prior iteration before starting a new one. This error indicates that your module has parameters that
were not used in producing loss. You can enable unused parameter detection by passing the keyword argument `find_unused_parameters=True` to `torch.nn.parallel
.DistributedDataParallel`, and by
making sure all `forward` function outputs participate in calculating loss.
If you already have done the above, then the distributed data parallel module wasn't able to locate the output tensors in the return value of your module's `f
orward` function. Please include the loss function and the structure of the return value of `forward` of your module when reporting this issue (e.g. list, dic
t, iterable).
Parameter indices which did not receive grad for rank 3: 1
In addition, you can set the environment variable TORCH_DISTRIBUTED_DEBUG to either INFO or DETAIL to print out information about which particular parameters
did not receive gradient on this rank as part of this error
Am I doing something wrong, or is the code intended to run only on a single GPU?
Do you happen to have a Pretrained HyperNetwork?
I looked it up, but I can't see it. Can I get it??
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 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.
An Open Source Machine Learning Framework for Everyone
Django
The Web framework for perfectionists with deadlines.
Laravel
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
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.