Dreambooth lora github txt caption files:. default="lora-dreambooth-model", help="The output directory where the model predictions and checkpoints will be written. {type} clothes, where {type} can be specified when adding a tag. ; PEFT/PETL method: how to update them. LoRA & Dreambooth training GUI & scripts preset & one key training environment for kohya-ss/sd-scripts NEW: Train WebUI The REAL Stable Diffusion Training Studio. Naive adaptation from 🤗Diffusers. FLUX, Stable Diffusion, SDXL, SD3, LoRA, Fine Tuning, DreamBooth, Training, Automatic1111, Forge WebUI, SwarmUI, DeepFake, TTS, Animation, Text To Video, Tutorials Implementation of "ZipLoRA: Any Subject in Any Style by Effectively Merging LoRAs" - mkshing/ziplora-pytorch Contribute to minghouse/train_dreambooth_lora development by creating an account on GitHub. You signed in with another tab or window. If this is still OOM, I'm wondering why should lora version exist? The next GPU memory capbility is 80G which can train it without lora. This notebook is designed to facilitate the training of Stable Diffusion 3 (SD3) models using the DreamBooth is a method to personalize text2image models like stable diffusion given just a few (3~5) images of a subject. py, you can find an example of how to use the model for inference. ; Run node main. main Hi, if you want help you will need to provide more info, we can't just guess what's your problem is. py script to train with LoRA. Once you are in, you need to Contribute to yardenfren1996/B-LoRA development by creating an account on GitHub. This notebook provides a boilerplate setup for training models using Dreambooth and LoRA with Hugging Face's Diffusers library in a cloud environment. This could be useful in e-commerce applications, for virtual try-on for example. DreamBooth is a method to personalize text2image models like stable diffusion given just a few(3~5) images of a subject. engine. lora finetune stable-diffusion dreambooth. github. Steps to reproduce the problem. enable LoRA for text encoder enable LoRA for U-Net In example. AI-powered developer platform Contribute to rupeshs/LoRA-DreamBooth-Training-UI-diffusers development by creating an account on GitHub. 🤗 Diffusers: State-of-the-art diffusion models for image, video, and audio generation in PyTorch and FLAX. - huggingface/diffusers Cog implementation of Huggingface Diffusers Dreambooth LoRA trainer - lucataco/cog-diffusers-dreambooth-lora FLUX, Stable Diffusion, SDXL, SD3, LoRA, Fine Tuning, DreamBooth, Training, Automatic1111, Forge WebUI, SwarmUI, DeepFake, TTS, Animation, Text To Video, Tutorials def validation_and_save(pretrained_model_path : str, transformer, validation_prompt : str, val_embeds, generator, output_folder, epoch, logger, global_step): Quantized training of Stable Diffusion 3 Medium to significantly reduce memory usage. Simplified cells to create the train_folder_directory and reg_folder_directory folders in kohya-dreambooth. Fine-tuning Stable Diffusion XL with DreamBooth and LoRA on a free-tier Colab Notebook 🧨 In this notebook, we show how to fine-tune Stable Diffusion XL (SDXL) with DreamBooth and LoRA dreambooth lora training. ; Run npm i to install dependencies. 0 model, unzip the mb_amg_gt_oue_dreambooth. Trainning Generative AI model. Reproduction. Installed version of pytorch_optimizer: 2. 0 as weights and 0. 5 months, finally I found best parameters for train model using by Dreambooth this ve Cog wrapper for Diffusers StableDiffusion3. License These are {repo_id} DreamBooth LoRA weights for {base_model}. ipynb for an example of how to merge Lora with Lora, and make inference dynamically using monkeypatch_add_lora. Unofficial implementation of DragDiffusion. Half Model - Enable this to save the model using half precision. 🤗 PEFT: State-of-the-art Parameter-Efficient Fine-Tuning. Top. utils. ; Loads a MusicGen checkpoint from the hub, for example the 1. I have resolved this issue with the following modifi I have been using train_dreambooth_lora_sdxl. If unsure go back to here; Place the {result}. By leveraging fine-tuning you You signed in with another tab or window. Dieses Repositoriy enthält die Google Colab Notebooks für folgende, drei Schritte: Notebook zur Erzeugen eines eigenen Datasets bestehend aus 9 Fotos einer He-Man Spielzeugfigur, welcher für das Feinetuning genutzt werden kann. Inpainting, simply put, it's a technique that allows to fill in missing parts of an image. Additionally, it includes modules for data preprocessing, model evaluation, and visualization of results. Raw. File metadata and controls. ) Automatic1111 Web UI - PC - Free. Enterprise-grade AI 🤗 Diffusers: State-of-the-art diffusion models for image, video, and audio generation in PyTorch and FLAX. 5B MusicGen Melody checkpoint. It save network as Lora, and may be merged in model back. validation_prompt` multiple times: `args. 5 ckpt. The train_dreambooth. safetensors file. 8 GB LoRA Training - Fix CUDA & xformers For DreamBooth and Textual Inversion in Automatic1111 SD UI. Since the original ImageBind model was not trained on some arbitrary number-naming Kolors is a large-scale text-to-image generation model based on latent diffusion, developed by the Kuaishou Kolors team. 5 faster on 4 GPUs as compared to 1 GPU. py script shows how to implement the training procedure and adapt it for stable diffusion. AI-powered developer platform train_dreambooth_lora_sdxl. But there are simply just so many parameters to set and which leads very bad results 1st parameter : Unfreeze Model we are already doing unet and text encoder tra Fine Tune DreamBooth. The LoRA training script is discussed in more detail in the LoRA Created with StackBlitz ⚡️. Hi, I'm trying to test training a dreambooth LoRA for SD1. In addition to some minor changes and rewording for clarity, this fork adds a slightly modified version of BLIP dataset autocaptioning functionality from victorchall's EveryDream comapnion tools repo Dreambooth LoRA Fine Tuning with Kohya-ss script. As a result, training is faster and it is easier to store the resulting weights because they are a lot smaller (~100MBs). 5-large-lora-trainer GitHub community articles Repositories. ps1 (windows) or run_gui. awesome - will try and report back! Thanks!! Saved searches Use saved searches to filter your results more quickly Fine-tuning "ImageBind One Embedding Space to Bind Them All" with LoRA - kyegomez/Gigabind GitHub is where people build software. ; If the result. Although LoRA was initially designed as a technique for reducing the number of trainable parameters in large-language models, the technique can also be applied to diffusion models. py script shows how to implement the training procedure with LoRA and adapt it for SANA. Use the train_dreambooth_lora. Make sure you have Cog installed. Loading. Logs. Loads an audio dataset using the datasets library, for example this small subset of songs in the punk style derived from the royalty-free PogChamp Music Classification Competition dataset. ; Notebook um aus dem eigenen Dataset mit Hilfe des Diffuser-Frameworks von Hugging Face ein eigenes LoRA zu trainieren. safetensors file can be used with the SDXL 1. - tengshaofeng/lora_tbq Notebooks using the Hugging Face libraries 🤗. Cog wrapper for Diffusers StableDiffusion3. import_utils import is_xformers_available 🤗 Diffusers: State-of-the-art diffusion models for image, video, and audio generation in PyTorch and FLAX. That model will appear on the left in the "model" dropdown. Regular dreambooth needs 100-200 steps per image to get it right (and maybe another 100-200 to get it perfect). Topics Trending Collections Enterprise Enterprise platform Cog implementation of the Diffusers Dreambooth LoRA Trainer. To do this, execute the This is a fork of the diffusers repository with the only difference being the addition of the train_dreambooth_inpaint_lora_sdxl. 0 create LoRA for Text Encoder: 72 modules. KaliYuga's simple fork of brian6091's LoRA-Enabled Dreambooth notebook. B-LoRA leverages the power of Stable Diffusion XL (SDXL) and Low-Rank Adaptation (LoRA) to disentangle the style and 🤗 Diffusers: State-of-the-art diffusion models for image, video, and audio generation in PyTorch and FLAX. I am using Dreambooth extension of automatic1111 that is what people love. example explanation: The dreambooth dataset contains the classes dog3, dog5, and dog8. And data is only one image. at this way the issue was resolved. Contribute to komojini/ComfyUI_SDXL_DreamBooth_LoRA_CustomNodes development by creating an account on GitHub. master This repository provides an engaging illustration on how to unleash the power of Stable Diffusion to finetune an inpainting model with your own images. lora finetune dreambooth Updated Aug 25, Thank you for your help! I tried your solution after watching video (runing commands to downgrade CUDA) but the problem remains. This script uses dreambooth technique, but with posibillity to train style via captions for all images (not just single concept). Improved the download link function from outside huggingface using parser. It's frustrating because I like the results I get training with regular Dreambooth more than training using Lora, so if I could do that quality of training, and then simplify it down to a Lora that would be perfect. Implementation of "ZipLoRA: Any Subject in Any Style by Effectively Merging LoRAs" - mkshing/ziplora-pytorch This repository contains the official implementation of the B-LoRA method, which enables implicit style-content separation of a single input image for various image stylization tasks. DreamBooth fine-tuning with LoRA. - FilippoO2/Quantized-Training-of-SD3 "Run dreambooth validation every X epochs. makes life much easier, less hyperparameter to tune (it just works!) Sometimes lora can lower the overall quality of the model, adjusting the adapter_weight or mixing it with other loras to balance can help Contribute to komojini/SDXL_DreamBooth_LoRA development by creating an account on GitHub. sh (linux) to start the GUI. (Optional) Generates Contribute to KaliYuga-ai/blip-lora-dreambooth-finetuning development by creating an account on GitHub. After about 2. Customizable tag layout for consistent tagging. Lora dreambooth training with inpainting tuned SDXL model - nikgli/train-lora-sdxl-inpaint Cog implementation of Huggingface Diffusers Dreambooth LoRA trainer - lucataco/cog-diffusers-dreambooth-lora Contribute to gdvstd/sd3-dreambooth-lora development by creating an account on GitHub. Train lora. 2334 lines (2334 loc) · 106 KB. models. I figured out StableDiffusionAttendAndExcitePipeline Contribute to PKU-ML/Diffusion-PID-Protection development by creating an account on GitHub. We have tested this script a lot of times and it works, so it can be literally anything. lora import LoRALinearLayer, text_encoder_lora_state_dict from diffusers. GitHub Gist: instantly share code, notes, and snippets. runtime. Before running the scripts, make sure to install the library's training dependencies: Important. something bug in save_model_hook? Reproduction accelerate launch train_ Contribute to d8ahazard/sd_dreambooth_extension development by creating an account on GitHub. training_utils import compute_snr, unet_lora_state_dict from diffusers. - huggingface/diffusers Before running the scripts, make sure to install the library's training dependencies: Important. js. From reddit on how to get a trained LORA to work with SD: "Create model" with the "source checkpoint" set to Stable Diffusion 1. Now let's get our dataset. Contribute to yardenfren1996/B-LoRA development by creating an account on GitHub. Thanks for your responses! I followed your suggestion, updated the code to the latest script version and made the changes to the two lines as AY-Liu recommended. lora create LoRA network. You are using a model of type clip_text_model to instantiate a fast-stable-diffusion Notebooks, A1111 + ComfyUI + DreamBooth Paperspace adaptations AUTOMATIC1111 Webui, ComfyUI and Dreambooth. 9. 5 as $\alpha$ . Now select your Lora model in the "Lora Model" Dropdown. Preview. It is pretty wired that even after I set lora rank to only 2, training the dreambooth_flux with lora still returns OOM. Contribute to gdvstd/sd3-dreambooth-lora development by creating an account on GitHub. Describe the bug When train_dreambooth_lora_flux attempts to generate images during validation, RuntimeError: Input type (float) and bias type (c10::BFloat16) should be the same is thrown Reproduction Just follow the steps from README_fl Clone this repository to your local machine. Stable Diffusion, SDXL, LoRA Training, DreamBooth Training, Automatic1111 Web UI, DeepFake, Deep Fakes, TTS, Animation, Text To Video, Tutorials, Guides, Lectures To use the PyTorch LoRA (Low-Rank Adaptation of Large Language Models) weights with the SDXL 1. Above results are from merging lora_illust. Updated Dec 17, 2024; Using Low-rank adaptation to quickly fine-tune diffusion models. json file next to main. py again with the merged source code, but I am still encountering an issue similar to #9237 during the log_validation stage. training_utils import unet_lora_state_dict sd3 dreambooth lora training book, adapted from the diffusers doc - Binxly/sd3-training Saved searches Use saved searches to filter your results more quickly Describe the bug I'm fine-tuning a model with Flux using LoRA and mixed precision (fp16). py). utils import check_min_version, is_wandb_available from diffusers. Reload to refresh your session. AI-powered developer platform kohya-LoRA-dreambooth. " This repo contains the Colab script to finetune Stable Diffusion XL(SDXL) model with DreamBooth and LoRA. DeepSpeedEngine'>. 10/19/2023 14:54:26 - INFO - __main__ - Distributed environment: NO Num processes: 1 Process index: 0 Local process index: 0 Device: cuda Mixed precision type: fp16 You are using a model of type clip_text_model to instantiate a model of type. Use lora. E. Implicit Style-Content Separation using B-LoRA. Learnable parameters: which layer or which module will be updated. - GitHub - dshly/SDXL-LoRA-Training: Dreambooth LoRA Fine Tuning with Kohya-ss script. Contribute to huchukato/dreambooth-lora-training-ui development by creating an account on GitHub. json file. 60 second sdxl Dreambooth lora training. base dim (rank): 8, alpha: 1. My GPU is three L40s (44G-48G). ipynb. Results in a smaller checkpoint with little noticeable difference in image output. pt files out of there. The resulting mb_amg_gt_oue_dreambooth and inclosed pytorch_lora_weights. LoRA & Dreambooth training GUI & scripts preset & one key training environment for kohya-ss/sd-scripts NEW: Train GUI Follow the installation guide below to install the GUI, then run run_gui. Enterprise-grade security features GitHub Copilot. 8. py is an end-to-end script that:. pt with lora_kiriko. The train_dreambooth_lora_sana. Topics Trending Collections Enterprise Enterprise platform. Contribute to Advocate99/DragDiffusion development by creating an account on GitHub. NA. You could use this script to fine-tune the SDXL inpainting model UNet via LoRA adaptation with your own subject images. Dreambooth (LoRA) with well-organized code structure. Would love any and all help! I'm using code from the commit agarwalml/kohya_ss@6c69b89 With one GPU, with around 1600 steps, I get a good DreamBooth is a method to personalize text2image models like stable diffusion given just a few(3~5) images of a subject. - huggingface/diffusers dreambooth lora training. io/) with the [SD3 diffusers The weights were trained on {prompt} using [DreamBooth] (https://dreambooth. The LR Scheduler settings allow you to control how LR changes during training. SDXL consists of a much larger UNet and two text encoders that make the cross-attention context quite larger than the previous variants. js and the file name was not changed, you should be able to just press enter. Same issue. Contribute to mindspore-lab/mindone development by creating an account on GitHub. e. py to train LoRA for specific character, It was working till like a week ago. . Since the original ImageBind model was not trained on some arbitrary number-naming LoRA is a training technique for significantly reducing the number of trainable parameters. 1. Placeholder tag templates: i. To try the original ImageBind model, set lora=False. py and convert_diffusers_sdxl_lora_to_webui. Paperspace Contribute to gdvstd/sd3-dreambooth-lora development by creating an account on GitHub. Footer Using Low-rank adaptation to quickly fine-tune diffusion models. pt with both 1. How To Do Stable Diffusion LORA Training By Using Web UI On Different Models - Tested SD 1. 12. add_argument("--peft_lora_path", default=None, type=str, required=True, help="Path to peft trained LoRA") finetune stable diffusion with Dreambooth、LoRA、ControlNet - WGS-note/finetune_stable_diffusion Lora Model - An existing lora checkpoint to load if resuming training, or to merge with the base model if generating a checkpoint. 0 base model. Contribute to nuwandda/sdxl-lora-training development by creating an account on GitHub. 5, SD 2. To make sure you can successfully run the latest versions of the example scripts, we highly recommend installing from source and keeping the install up to date as we update the example scripts frequently and install some example-specific requirements. Candy Machine is a nascent image tagger for manually tagging small datasets (< 1k images) with . ; Training process: default to diffuion-denoising, which can be extended like XTI. You switched accounts on another tab or window. ) Automatic1111 Web UI - PC - from diffusers. 0 [Dreambooth] pytorch_optimizer camenduru/peft-lora-sd-dreambooth-colab This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository. Updated Dec 1, 2024; In example. Contribute to huggingface/notebooks development by creating an account on GitHub. As SD3 is gated, before using it with diffusers you first need to go to the Stable Diffusion 3 Medium Hugging Face page, fill in the form and accept the gate. py: Dreambooth training script path: RESOLUTION: 1024: Resolution of the images: MAX_TRAIN_STEPS: 500: Total number of training steps: This readme file will get updated if be necessary so always checkout this file if something not working and open an issue thread on our GitHub repo; How to download and install DreamBooth extension for Automatic1111 Web UI; How to train with DreamBooth extension for Studio Photoshoot realism; What are class images and why do we use them LoRa uses a separate set of Learning Rate fields because the LR values are much higher for LoRa than normal dreambooth. To do this, execute the 🤗 Diffusers: State-of-the-art diffusion models for image, video, and audio generation in PyTorch and FLAX. The script dreambooth_musicgen. json file is located next to main. No 🤗 Diffusers: State-of-the-art diffusion models for image, video, and audio generation in PyTorch and FLAX. - huggingface/diffusers "Run dreambooth validation every X epochs. Fine-tuning Stable Diffusion XL with DreamBooth and LoRA on a T4 GPU. Scripts to tune SD models using DreamBooth and Diffusers - vltmedia/DreamBoothTune Contribute to komojini/SDXL_DreamBooth_LoRA development by creating an account on GitHub. ; It allows we conduct a unified training pipeline with This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository. Now, during the forward pass, I'm encountering a new iss. GitHub is where people build software. py. LoRA & Dreambooth training scripts & GUI use kohya-ss's trainer, for diffusion model. The weights were trained using [DreamBooth](https://dreambooth. You signed out in another tab or window. ipynb and kohya-LoRA-dreambooth. - huggingface/diffusers 25+ Stable Diffusion Tutorials - Guides - DreamBooth - Textual Inversion - LoRA - ControlNet - Style Transfer - RunPod - Animation Removed the download and generate regularization images function from kohya-dreambooth. Dreambooth validation consists of running the prompt" " `args. Checkout scripts/merge_lora_with_lora. This will also allow us to push the trained model parameters to the Hugging Face Hub platform. This project used Dreambooth+LoRA to generate customized model for your own pets, and uses AnimateDiff to animate your pet - rorschach-xiao/PetAvatar Lora Model - An existing lora checkpoint to load if resuming training, or to merge with the base model if generating a checkpoint. Yet, i Stable Diffusion XL LoRa Training. (If it doesn't exist, put your Lora PT file here: Automatic1111\stable-diffusion-webui\models Describe the bug I tried running train_dreambooth_lora_flux. - huggingface/diffusers Web-Based Manual Image Tagger for Training Custom Stable Diffusion LORAs & DreamBooth Models. GitHub community articles Repositories. Commit and libraries. - cloneofsimo/lora In UniDiffusion, all training methods are decomposed into three dimensions. ",) parser I am going to try to keep this as short as possible, if you want more "in depth" information, there are many tutorials and guides online that explain Stable Diffusion LoRA/Dreambooth training in greater detail. Code. from diffusers. For LoRa , the LR defaults are 1e-4 for UNET and 5e-5 for Text. " Contribute to komojini/comfyui-sdxl-dreambooth-lora development by creating an account on GitHub. create LoRA for U-Net: 192 modules. Generate Lora weights for extra networks will save a file in the model/lora directory that will work (as long as you don't use extended Lora). In Octoparse, start the scraping process and extract data as a . The weights were trained using [🧨 diffusers Advanced Dreambooth Training Script](https://github. I recommend deleting or moving all the old incompatible . Blame. Saved searches Use saved searches to filter your results more quickly Contribute to minghouse/train_dreambooth_lora development by creating an account on GitHub. Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, Rank 128 works best for me with around 1500 training steps--optimizer="prodigy" with --learning_rate=1. This guide demonstrates how to use LoRA, a low-rank approximation technique, to fine-tune DreamBooth with the CompVis/stable-diffusion-v1-4 model. It raise ValueError: unexpected save model: <class 'deepspeed. Advanced Security. py script. Describe the bug I am trying to run the famous colab notebook SDXL_DreamBooth_LoRA_. num_validation_images`. g. I am using the same baseline model and the same data. I realized that previous size of all the LoRA files had 29967176 bytes, now it has 29889672 and less keys in dict after I load it as pure . Trained on billions of text-image pairs, Kolors exhibits significant advantages over both open-source and closed-source models in visual quality, complex semantic accuracy, and text rendering for both Chinese and English characters. zip file. - NSTiwari/Stable-DiffusionXL-using-DreamBooth-and-LoRA GitHub community articles Repositories. com/huggingface/diffusers/blob/main/examples/advanced_diffusion_training/train_dreambooth_lora_sdxl_advanced. AI-powered developer platform Available add-ons. How to use. With the LORA thing checked, after 1000 steps per image it's not even close (i can see it's getting there, it has an idea of what's going on, but the results are like with DB after 20-50 steps). The model was fine-tuned with approximately 20 images, each You signed in with another tab or window. This is not supported for all configurations of models and can yield errors. If the result is generated using dreambooth_lora, it makes sense; however, if it is not used and only sd1. DreamBooth is a method to personalize text-to-image models like flux, stable diffusion given just a few (3~5) images of a subject. 5-Large LoRA Training - lucataco/cog-stable-diffusion-3. Download images from here and save them Hey everybody! 😇 I am working on find best parameters for training my stable diffusion model with my face in different ways with 16 VRAM. It's so weird. More than 100 million people use GitHub to discover, fork, and contribute to over 420 million projects. 5 is employed, then the generated result is meaningless. To try the LoRA fine-tuned model, change lora=True within the script. Contribute to d8ahazard/sd_dreambooth_extension development by creating an account on GitHub. Implementation of "ZipLoRA: Any Subject in Any Style by Effectively Merging LoRAs" - mkshing/ziplora-pytorch @inproceedings{ruiz2023dreambooth, title={Dreambooth: Fine tuning text-to-image diffusion models for subject-driven generation}, author={Ruiz, Nataniel and Li, Yuanzhen and Jampani, Varun and Pritch, Yael and Rubinstein, Michael and Aberman, Kfir}, booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition I try some of the saved lora files that it made during training. optimization import get_scheduler from diffusers. Saved searches Use saved searches to filter your results more quickly Describe the bug when I run the script train_dreambooth_lora_flux. - huggingface/peft Hi, I am wondering how to combine Attend and Excite pipeline with DreamBooth+LoRA / Custom Diffusion setting. DreamBooth is a method to personalize text2image models like stable diffusion given just a few (3~5) images of a subject. In-built image FLUX, Stable Diffusion, SDXL, SD3, LoRA, Fine Tuning, DreamBooth, Training, Automatic1111, Forge WebUI, SwarmUI, DeepFake, TTS, Animation, Text To Video, Tutorials SANA: Efficient High-Resolution Image Synthesis with Linear Diffusion Transformer - NVlabs/Sana Whether I included the Lora or not, it gave pretty much the same image. Initially, I ran into an out-of-memory error, which I solved by upgrading to a larger GPU. Contribute to imvamoss/dreambooth_lora_car development by creating an account on GitHub. ipynb to build a dreambooth model out of sdxl + vae using accelerate launch train_dreambooth_lora_sdxl. , finetune, low-rank adaption, adapter, etc. Stable Diffusion XL fine-tuning with Dreambooth & Lora: how to structure local dataset for fine-tuning with ROI import network module: networks. Maybe there's something wrong with my Stable Diffusion or DreamBooth environment, but still thank you very much. - huggingface/diffusers one for all, Optimal generator with No Exception. LoRA - Low-Rank Adaption of Large Language Models, was first introduced by Microsoft in LoRA: Low-Rank Adaptation of Large Language Models by Edward J. Contribute to lucifertrj/Dreambooth-LoRA development by creating an account on GitHub. default="lora-dreambooth-model", help="The output directory where the model predictions and checkpoints will be written This project is an implementation of fine-tuning an SDXL model using DreamBooth and LoRA on custom data of interior rooms to generate designs for your home. - huggingface/diffusers 🤗 Diffusers: State-of-the-art diffusion models for image, video, and audio generation in PyTorch and FLAX. io/). onfk tlk txee vci nguoan abcb skysr ayxsawy zksx dhb