Snapshot của XavierXiao/Dreambooth-Stable-Diffusion: 7.7k★ · Jupyter Notebook · AI Tools. Implementation of Dreambooth (https://arxiv.org/abs/2208.12242) with Stable Diffusion
Tóm tắt dựng từ metadata GitHub của chính dự án — chưa có bài review TopGit. Trang sẽ tự động cập nhật khi bài review đầy đủ được xuất bản.
VÌ SAO CHƯA CÓ REVIEW
TopGit viết bài đầy đủ cho repo có nhiều sao nhất và được yêu cầu nhiều nhất. Trang này là snapshot trong thời gian chờ — xem README gốc ở tab READ ME.
This is an implementtaion of Google's Dreambooth with Stable Diffusion. The original Dreambooth is based on Imagen text-to-image model. However, neither the model nor the pre-trained weights of Imagen is available. To enable people to fine-tune a text-to-image model with a few examples, I implemented the idea of Dreambooth on Stable diffusion.
This code repository is based on that of Textual Inversion. Note that Textual Inversion only optimizes word ebedding, while dreambooth fine-tunes the whole diffusion model.
The implementation makes minimum changes over the official codebase of Textual Inversion. In fact, due to lazyness, some components in Textual Inversion, such as the embedding manager, are not deleted, although they will never be used here.
Update
9/20/2022: I just found a way to reduce the GPU memory a bit. Remember that this code is based on Textual Inversion, and TI's code base has this line, which disable gradient checkpointing in a hard-code way. This is because in TI, the Unet is not optimized. However, in Dreambooth we optimize the Unet, so we can turn on the gradient checkpoint pointing trick, as in the original SD repo here. The gradient checkpoint is default to be True in config. I have updated the codes.
Usage
Preparation
First set-up the ldm enviroment following the instruction from textual inversion repo, or the original Stable Diffusion repo.
To fine-tune a stable diffusion model, you need to obtain the pre-trained stable diffusion models following their instructions. Weights can be downloaded on HuggingFace. You can decide which version of checkpoint to use, but I use sd-v1-4-full-ema.ckpt.
We also need to create a set of images for regularization, as the fine-tuning algorithm of Dreambooth requires that. Details of the algorithm can be found in the paper. Note that in the original paper, the regularization images seem to be generated on-the-fly. However, here I generated a set of regularization images before the training. The text prompt for generating regularization images can be photo of a <class>, where <class> is a word that describes the class of your object, such as dog. The command is
python scripts/stable_txt2img.py --ddim_eta 0.0 --n_samples 8 --n_iter 1 --scale 10.0 --ddim_steps 50 --ckpt /path/to/original/stable-diffusion/sd-v1-4-full-ema.ckpt --prompt "a photo of a <class>"
I generate 8 images for regularization, but more regularization images may lead to stronger regularization and better editability. After that, save the generated images (separately, one image per .png file) at /root/to/regularization/images.
Updates on 9/9
We should definitely use more images for regularization. Please try 100 or 200, to better align with the original paper. To acomodate this, I shorten the "repeat" of reg dataset in the config file.
For some cases, if the generated regularization images are highly unrealistic (happens when you want to generate "man" or "woman"), you can find a diverse set of images (of man/woman) online, and use them as regularization images.
Training
Training can be done by running the following command
Detailed configuration can be found in configs/stable-diffusion/v1-finetune_unfrozen.yaml. In particular, the default learning rate is 1.0e-6 as I found the 1.0e-5 in the Dreambooth paper leads to poor editability. The parameter reg_weight corresponds to the weight of regularization in the Dreambooth paper, and the default is set to 1.0.
Dreambooth requires a placeholder word [V], called identifier, as in the paper. This identifier needs to be a relatively rare tokens in the vocabulary. The original paper approaches this by using a rare word in T5-XXL tokenizer. For simplicity, here I just use a random word sks and hard coded it.. If you want to change that, simply make a change in this file.
Training will be run for 800 steps, and two checkpoints will be saved at ./logs/<job_name>/checkpoints, one at 500 steps and one at final step. Typically the one at 500 steps works well enough. I train the model use two A6000 GPUs and it takes ~15 mins.
Generation
After training, personalized samples can be obtained by running the command
python scripts/stable_txt2img.py --ddim_eta 0.0
--n_samples 8
--n_iter 1
--scale 10.0
--ddim_steps 100
--ckpt /path/to/saved/checkpoint/from/training
--prompt "photo of a sks <class>"
In particular, sks is the identifier, which should be replaced by your choice if you happen to change the identifier, and <class> is the class word --class_word for training.
Results
Here I show some qualitative results. The training images are obtained from the issue in the Textual Inversion repository, and they are 3 images of a large trash container. Regularization images are generated by prompt photo of a container. Regularization images are shown here:
After training, generated images with prompt photo of a sks container:
Generated images with prompt photo of a sks container on the beach:
Generated images with prompt photo of a sks container on the moon:
Some not-so-perfect but still interesting results:
Generated images with prompt photo of a red sks container:
Generated images with prompt a dog on top of sks container:
XavierXiao/Dreambooth-Stable-Diffusion thuộc nhóm AI Tools trên TopGit, cùng 4 topic GitHub. Trang Trending và Topics liệt kê các repo cùng số sao và cùng ngôn ngữ để so sánh.
Đọc thêm về XavierXiao/Dreambooth-Stable-Diffusion ở đâu?
Trang TopGit này là một snapshot — tab "Readme" hiển thị nguyên văn README của repo (đã bỏ link, giữ ảnh). Repo GitHub ở github.com/XavierXiao/Dreambooth-Stable-Diffusion là nguồn chính thức.
Vì sao XavierXiao/Dreambooth-Stable-Diffusion được xếp vào nhóm AI Tools?
TopGit xếp XavierXiao/Dreambooth-Stable-Diffusion vào nhóm AI Tools dựa trên GitHub topics và mô tả của repo (gắn thẻ: "pytorch", "pytorch-lightning", "stable-diffusion"). Việc phân loại dựa trên metadata thật của repo, không phải đoán theo cảm tính biên tập.
XavierXiao/Dreambooth-Stable-Diffusion có phải mã nguồn mở không?
Có — XavierXiao/Dreambooth-Stable-Diffusion phát hành theo license MIT, nghĩa là mã nguồn mở để đọc, fork và (tùy license) tái sử dụng. Mã: github.com/XavierXiao/Dreambooth-Stable-Diffusion.
XavierXiao/Dreambooth-Stable-Diffusion có website riêng không?
TopGit chưa ghi nhận URL trang chủ cho XavierXiao/Dreambooth-Stable-Diffusion. Phần README ở tab phía trên thường có link demo, hoặc xem mô tả GitHub của repo.
XavierXiao/Dreambooth-Stable-Diffusion dùng license gì?
XavierXiao/Dreambooth-Stable-Diffusion phát hành theo license MIT. Nên mở file LICENSE trên GitHub để xác nhận — license metadata đôi khi lệch với thực tế dự án.
XavierXiao/Dreambooth-Stable-Diffusion là gì?
XavierXiao/Dreambooth-Stable-Diffusion (XavierXiao/Dreambooth-Stable-Diffusion) là dự án Jupyter Notebook trên GitHub. Theo mô tả gốc: Implementation of Dreambooth (https://arxiv.org/abs/2208.12242) with Stable Diffusion
Đọc đầy đủ README ở tab phía trên.
Muốn nghe thêm một ý kiến về Dreambooth-Stable-Diffusion?
Hỏi một AI đọc được trang này — một cú bấm là có ngay nhận định về Dreambooth-Stable-Diffusion.