__SDXL is now supported. The sdxl branch has been merged into the main branch. If you update the repository, please follow the upgrade instructions. Also, the version of accelerate has been updated, so please run accelerate config again.__ The documentation for SDXL training is [here](./README.md#sdxl-training).
This repository contains training, generation and utility scripts for Stable Diffusion.
[__Change History__](#change-history) is moved to the bottom of the page.
[__Change History__](#change-history) is moved to the bottom of the page.
更新履歴は[ページ末尾](#change-history)に移しました。
[日本語版README](./README-ja.md)
[日本語版READMEはこちら](./README-ja.md)
For easier use (GUI and PowerShell scripts etc...), please visit [the repository maintained by bmaltais](https://github.com/bmaltais/kohya_ss). Thanks to @bmaltais!
@@ -16,142 +18,13 @@ This repository contains the scripts for:
* Image generation
* Model conversion (supports 1.x and 2.x, Stable Diffision ckpt/safetensors and Diffusers)
__Stable Diffusion web UI now seems to support LoRA trained by ``sd-scripts``.__ Thank you for great work!!!
## About SDXL training
The feature of SDXL training is now available in sdxl branch as an experimental feature.
Sep 24, 2023: The feature will be merged into the main branch very soon. Following are the changes from the previous version.
-`accelerate` is updated to 0.23.0, and `diffusers` is updated to 0.21.2. Dependency for `invisible-watermark` is removed. Please update them with the upgrade instructions below.
- Intel ARC support with IPEX is added. [#825](https://github.com/kohya-ss/sd-scripts/pull/825)
- Other changes and fixes.
- Thanks for contributions from Disty0, sdbds, jvkap, rockerBOO, Symbiomatrix and others!
Sep 3, 2023:
- ControlNet-LLLite is added. See [documentation](./docs/train_lllite_README.md) for details.
- JPEG XL is supported. [#786](https://github.com/kohya-ss/sd-scripts/pull/786)
- Peak memory usage is reduced. [#791](https://github.com/kohya-ss/sd-scripts/pull/791)
- Input perturbation noise is added. See [#798](https://github.com/kohya-ss/sd-scripts/pull/798) for details.
- Dataset subset now has `caption_prefix` and `caption_suffix` options. The strings are added to the beginning and the end of the captions before shuffling. You can specify the options in `.toml`.
- Other minor changes.
- Thanks for contributions from Isotr0py, vvern999, lansing and others!
Aug 13, 2023:
- LoRA-FA is added experimentally. Specify `--network_module networks.lora_fa` option instead of `--network_module networks.lora`. The trained model can be used as a normal LoRA model.
Aug 12, 2023:
- The default value of noise offset when omitted has been changed to 0 from 0.0357.
- The different learning rates for each U-Net block are now supported. Specify with `--block_lr` option. Specify 23 values separated by commas like `--block_lr 1e-3,1e-3 ... 1e-3`.
- [SAI Model Spec](https://github.com/Stability-AI/ModelSpec) metadata is now supported partially. `hash_sha256` is not supported yet.
- The main items are set automatically.
- You can set title, author, description, license and tags with `--metadata_xxx` options in each training script.
- Merging scripts also support minimum SAI Model Spec metadata. See the help message for the usage.
- Metadata editor will be available soon.
- SDXL LoRA has `sdxl_base_v1-0` now for `ss_base_model_version` metadata item, instead of `v0-9`.
Aug 4, 2023:
-`bitsandbytes` is now optional. Please install it if you want to use it. The insructions are in the later section.
-`albumentations` is not required anymore.
- An issue for pooled output for Textual Inversion training is fixed.
-`--v_pred_like_loss ratio` option is added. This option adds the loss like v-prediction loss in SDXL training. `0.1` means that the loss is added 10% of the v-prediction loss. The default value is None (disabled).
- In v-prediction, the loss is higher in the early timesteps (near the noise). This option can be used to increase the loss in the early timesteps.
- Arbitrary options can be used for Diffusers' schedulers. For example `--lr_scheduler_args "lr_end=1e-8"`.
-`sdxl_gen_imgs.py` supports batch size > 1.
- Fix ControlNet to work with attention couple and reginal LoRA in `gen_img_diffusers.py`.
Summary of the feature:
-`tools/cache_latents.py` is added. This script can be used to cache the latents to disk in advance.
- The options are almost the same as `sdxl_train.py'. See the help message for the usage.
- This script should work with multi-GPU, but it is not tested in my environment.
-`tools/cache_text_encoder_outputs.py` is added. This script can be used to cache the text encoder outputs to disk in advance.
- The options are almost the same as `cache_latents.py' and `sdxl_train.py'. See the help message for the usage.
-`sdxl_train.py` is a script for SDXL fine-tuning. The usage is almost the same as `fine_tune.py`, but it also supports DreamBooth dataset.
-`--full_bf16` option is added. Thanks to KohakuBlueleaf!
- This option enables the full bfloat16 training (includes gradients). This option is useful to reduce the GPU memory usage.
- However, bitsandbytes==0.35 doesn't seem to support this. Please use a newer version of bitsandbytes or another optimizer.
- I cannot find bitsandbytes>0.35.0 that works correctly on Windows.
- In addition, the full bfloat16 training might be unstable. Please use it at your own risk.
-`prepare_buckets_latents.py` now supports SDXL fine-tuning.
-`sdxl_train_network.py` is a script for LoRA training for SDXL. The usage is almost the same as `train_network.py`.
- Both scripts has following additional options:
-`--cache_text_encoder_outputs` and `--cache_text_encoder_outputs_to_disk`: Cache the outputs of the text encoders. This option is useful to reduce the GPU memory usage. This option cannot be used with options for shuffling or dropping the captions.
-`--no_half_vae`: Disable the half-precision (mixed-precision) VAE. VAE for SDXL seems to produce NaNs in some cases. This option is useful to avoid the NaNs.
- The image generation during training is now available. `--no_half_vae` option also works to avoid black images.
-`--weighted_captions` option is not supported yet for both scripts.
-`--min_timestep` and `--max_timestep` options are added to each training script. These options can be used to train U-Net with different timesteps. The default values are 0 and 1000.
-`sdxl_train_textual_inversion.py` is a script for Textual Inversion training for SDXL. The usage is almost the same as `train_textual_inversion.py`.
-`--cache_text_encoder_outputs` is not supported.
-`token_string` must be alphabet only currently, due to the limitation of the open-clip tokenizer.
- There are two options for captions:
1. Training with captions. All captions must include the token string. The token string is replaced with multiple tokens.
2. Use `--use_object_template` or `--use_style_template` option. The captions are generated from the template. The existing captions are ignored.
- See below for the format of the embeddings.
-`sdxl_gen_img.py` is added. This script can be used to generate images with SDXL, including LoRA. See the help message for the usage.
- Textual Inversion is supported, but the name for the embeds in the caption becomes alphabet only. For example, `neg_hand_v1.safetensors` can be activated with `neghandv`.
`requirements.txt` is updated to support SDXL training.
### Tips for SDXL training
- The default resolution of SDXL is 1024x1024.
- The fine-tuning can be done with 24GB GPU memory with the batch size of 1. For 24GB GPU, the following options are recommended __for the fine-tuning with 24GB GPU memory__:
- Train U-Net only.
- Use gradient checkpointing.
- Use `--cache_text_encoder_outputs` option and caching latents.
- Use Adafactor optimizer. RMSprop 8bit or Adagrad 8bit may work. AdamW 8bit doesn't seem to work.
- The LoRA training can be done with 8GB GPU memory (10GB recommended). For reducing the GPU memory usage, the following options are recommended:
- Train U-Net only.
- Use gradient checkpointing.
- Use `--cache_text_encoder_outputs` option and caching latents.
- Use one of 8bit optimizers or Adafactor optimizer.
- Use lower dim (-8 for 8GB GPU).
-`--network_train_unet_only` option is highly recommended for SDXL LoRA. Because SDXL has two text encoders, the result of the training will be unexpected.
- PyTorch 2 seems to use slightly less GPU memory than PyTorch 1.
-`--bucket_reso_steps` can be set to 32 instead of the default value 64. Smaller values than 32 will not work for SDXL training.
Example of the optimizer settings for Adafactor with the fixed learning rate:
These files do not contain requirements for PyTorch. Because the versions of them depend on your environment. Please install PyTorch at first (see installation guide below.)
The scripts are tested with PyTorch 1.12.1 and 2.0.1, Diffusers 0.18.2.
The scripts are tested with Pytorch 2.0.1. 1.12.1 is not tested but should work.
The documentation in this section will be moved to a separate document later.
### Training scripts for SDXL
- `sdxl_train.py` is a script for SDXL fine-tuning. The usage is almost the same as `fine_tune.py`, but it also supports DreamBooth dataset.
- `--full_bf16` option is added. Thanks to KohakuBlueleaf!
- This option enables the full bfloat16 training (includes gradients). This option is useful to reduce the GPU memory usage.
- The full bfloat16 training might be unstable. Please use it at your own risk.
- The different learning rates for each U-Net block are now supported in sdxl_train.py. Specify with `--block_lr` option. Specify 23 values separated by commas like `--block_lr 1e-3,1e-3 ... 1e-3`.
- `prepare_buckets_latents.py` now supports SDXL fine-tuning.
- `sdxl_train_network.py` is a script for LoRA training for SDXL. The usage is almost the same as `train_network.py`.
- Both scripts has following additional options:
- `--cache_text_encoder_outputs` and `--cache_text_encoder_outputs_to_disk`: Cache the outputs of the text encoders. This option is useful to reduce the GPU memory usage. This option cannot be used with options for shuffling or dropping the captions.
- `--no_half_vae`: Disable the half-precision (mixed-precision) VAE. VAE for SDXL seems to produce NaNs in some cases. This option is useful to avoid the NaNs.
- `--weighted_captions` option is not supported yet for both scripts.
- `sdxl_train_textual_inversion.py` is a script for Textual Inversion training for SDXL. The usage is almost the same as `train_textual_inversion.py`.
- `--cache_text_encoder_outputs` is not supported.
- There are two options for captions:
1. Training with captions. All captions must include the token string. The token string is replaced with multiple tokens.
2. Use `--use_object_template` or `--use_style_template` option. The captions are generated from the template. The existing captions are ignored.
- See below for the format of the embeddings.
- `--min_timestep` and `--max_timestep` options are added to each training script. These options can be used to train U-Net with different timesteps. The default values are 0 and 1000.
### Utility scripts for SDXL
- `tools/cache_latents.py` is added. This script can be used to cache the latents to disk in advance.
- The options are almost the same as `sdxl_train.py'. See the help message for the usage.
- This script should work with multi-GPU, but it is not tested in my environment.
-`tools/cache_text_encoder_outputs.py` is added. This script can be used to cache the text encoder outputs to disk in advance.
- The options are almost the same as `cache_latents.py` and `sdxl_train.py`. See the help message for the usage.
-`sdxl_gen_img.py` is added. This script can be used to generate images with SDXL, including LoRA, Textual Inversion and ControlNet-LLLite. See the help message for the usage.
### Tips for SDXL training
- The default resolution of SDXL is 1024x1024.
- The fine-tuning can be done with 24GB GPU memory with the batch size of 1. For 24GB GPU, the following options are recommended __for the fine-tuning with 24GB GPU memory__:
- Train U-Net only.
- Use gradient checkpointing.
- Use `--cache_text_encoder_outputs` option and caching latents.
- Use Adafactor optimizer. RMSprop 8bit or Adagrad 8bit may work. AdamW 8bit doesn't seem to work.
- The LoRA training can be done with 8GB GPU memory (10GB recommended). For reducing the GPU memory usage, the following options are recommended:
- Train U-Net only.
- Use gradient checkpointing.
- Use `--cache_text_encoder_outputs` option and caching latents.
- Use one of 8bit optimizers or Adafactor optimizer.
- Use lower dim (4 to 8 for 8GB GPU).
-`--network_train_unet_only` option is highly recommended for SDXL LoRA. Because SDXL has two text encoders, the result of the training will be unexpected.
- PyTorch 2 seems to use slightly less GPU memory than PyTorch 1.
-`--bucket_reso_steps` can be set to 32 instead of the default value 64. Smaller values than 32 will not work for SDXL training.
Example of the optimizer settings for Adafactor with the fixed learning rate:
ControlNet-LLLite, a novel method for ControlNet with SDXL, is added. See [documentation](./docs/train_lllite_README.md) for details.
## Change History
### 15 Jun. 2023, 2023/06/15
### Oct 1. 2023 / 2023/10/1
- Prodigy optimizer is supported in each training script. It is a member of D-Adaptation and is effective for DyLoRA training. [PR #585](https://github.com/kohya-ss/sd-scripts/pull/585) Please see the PR for details. Thanks to sdbds!
- Install the package with `pip install prodigyopt`. Then specify the option like `--optimizer_type="prodigy"`.
- Arbitrary Dataset is supported in each training script (except XTI). You can use it by defining a Dataset class that returns images and captions.
- Prepare a Python script and define a class that inherits `train_util.MinimalDataset`. Then specify the option like `--dataset_class package.module.DatasetClass` in each training script.
- Please refer to `MinimalDataset` for implementation. I will prepare a sample later.
- The following features have been added to the generation script.
- Added an option `--highres_fix_disable_control_net` to disable ControlNet in the 2nd stage of Highres. Fix. Please try it if the image is disturbed by some ControlNet such as Canny.
- Added Variants similar to sd-dynamic-propmpts in the prompt.
- If you specify `{spring|summer|autumn|winter}`, one of them will be randomly selected.
- If you specify `{2$$chocolate|vanilla|strawberry}`, two of them will be randomly selected.
- If you specify `{1-2$$ and $$chocolate|vanilla|strawberry}`, one or two of them will be randomly selected and connected by ` and `.
- You can specify the number of candidates in the range `0-2`. You cannot omit one side like `-2` or `1-`.
- It can also be specified for the prompt option.
- If you specify `e` or `E`, all candidates will be selected and the prompt will be repeated multiple times (`--images_per_prompt` is ignored). It may be useful for creating X/Y plots.
- You can also specify `--am {e$$0.2|0.4|0.6|0.8|1.0},{e$$0.4|0.7|1.0} --d 1234`. In this case, 15 prompts will be generated with 5*3.
- There is no weighting function.
- SDXL training is now available in the main branch. The sdxl branch is merged into the main branch.
- [SAI Model Spec](https://github.com/Stability-AI/ModelSpec) metadata is now supported partially. `hash_sha256` is not supported yet.
- The main items are set automatically.
- You can set title, author, description, license and tags with `--metadata_xxx` options in each training script.
- Merging scripts also support minimum SAI Model Spec metadata. See the help message for the usage.
- Metadata editor will be available soon.
### 8 Jun. 2023, 2023/06/08
-`bitsandbytes` is now optional. Please install it if you want to use it. The insructions are in the later section.
- Fixed a bug where clip skip did not work when training with weighted captions (`--weighted_captions` specified) and when generating sample images during training.
-`--v_pred_like_loss ratio` option is added. This option adds the loss like v-prediction loss in SDXL training. `0.1` means that the loss is added 10% of the v-prediction loss. The default value is None (disabled).
- In v-prediction, the loss is higher in the early timesteps (near the noise). This option can be used to increase the loss in the early timesteps.
- Fix `train_network.py` to probably work with older versions of LyCORIS.
- `gen_img_diffusers.py` now supports `BREAK` syntax.
- Arbitrary options can be used for Diffusers' schedulers. For example `--lr_scheduler_args "lr_end=1e-8"`.
### 3 Jun. 2023, 2023/06/03
- LoRA-FA is added experimentally. Specify `--network_module networks.lora_fa` option instead of `--network_module networks.lora`. The trained model can be used as a normal LoRA model.
- JPEG XL is supported. [#786](https://github.com/kohya-ss/sd-scripts/pull/786)
- Input perturbation noise is added. See [#798](https://github.com/kohya-ss/sd-scripts/pull/798) for details.
- Dataset subset now has `caption_prefix` and `caption_suffix` options. The strings are added to the beginning and the end of the captions before shuffling. You can specify the options in `.toml`.
- Intel ARC support with IPEX is added. [#825](https://github.com/kohya-ss/sd-scripts/pull/825)
- Other bug fixes and improvements.
- Max Norm Regularization is now available in `train_network.py`. [PR #545](https://github.com/kohya-ss/sd-scripts/pull/545) Thanks to AI-Casanova!
- Max Norm Regularization is a technique to stabilize network training by limiting the norm of network weights. It may be effective in suppressing overfitting of LoRA and improving stability when used with other LoRAs. See PR for details.
- Specify as `--scale_weight_norms=1.0`. It seems good to try from `1.0`.
- The networks other than LoRA in this repository (such as LyCORIS) do not support this option.
- Three types of dropout have been added to `train_network.py` and LoRA network.
- Dropout is a technique to suppress overfitting and improve network performance by randomly setting some of the network outputs to 0.
- `--network_dropout` is a normal dropout at the neuron level. In the case of LoRA, it is applied to the output of down. Proposed in [PR #545](https://github.com/kohya-ss/sd-scripts/pull/545) Thanks to AI-Casanova!
- `--network_dropout=0.1` specifies the dropout probability to `0.1`.
- Note that the specification method is different from LyCORIS.
- For LoRA network, `--network_args` can specify `rank_dropout` to dropout each rank with specified probability. Also `module_dropout` can be specified to dropout each module with specified probability.
- Specify as `--network_args "rank_dropout=0.2" "module_dropout=0.1"`.
- `--network_dropout`, `rank_dropout`, and `module_dropout` can be specified at the same time.
- Values of 0.1 to 0.3 may be good to try. Values greater than 0.5 should not be specified.
- `rank_dropout` and `module_dropout` are original techniques of this repository. Their effectiveness has not been verified yet.
- The networks other than LoRA in this repository (such as LyCORIS) do not support these options.
- Added an option `--scale_v_pred_loss_like_noise_pred` to scale v-prediction loss like noise prediction in each training script.
- By scaling the loss according to the time step, the weights of global noise prediction and local noise prediction become the same, and the improvement of details may be expected.
- See [this article](https://xrg.hatenablog.com/entry/2023/06/02/202418) by xrg for details (written in Japanese). Thanks to xrg for the great suggestion!
- Max Norm Regularizationが`train_network.py`で使えるようになりました。[PR #545](https://github.com/kohya-ss/sd-scripts/pull/545) AI-Casanova氏に感謝します。
- Max Norm Regularizationは、ネットワークの重みのノルムを制限することで、ネットワークの学習を安定させる手法です。LoRAの過学習の抑制、他のLoRAと併用した時の安定性の向上が期待できるかもしれません。詳細はPRを参照してください。
- Show warning when image caption file does not exist during training. [PR #533](https://github.com/kohya-ss/sd-scripts/pull/533) Thanks to TingTingin!
- Warning is also displayed when using class+identifier dataset. Please ignore if it is intended.
- `train_network.py` now supports merging network weights before training. [PR #542](https://github.com/kohya-ss/sd-scripts/pull/542) Thanks to u-haru!
- `--base_weights` option specifies LoRA or other model files (multiple files are allowed) to merge.
- `--base_weights_multiplier` option specifies multiplier of the weights to merge (multiple values are allowed). If omitted or less than `base_weights`, 1.0 is used.
- This is useful for incremental learning. See PR for details.
- Show warning and continue training when uploading to HuggingFace fails.
- [D-Adaptation v3.0](https://github.com/facebookresearch/dadaptation) is now supported. [PR #530](https://github.com/kohya-ss/sd-scripts/pull/530) Thanks to sdbds!
- `--optimizer_type` now accepts `DAdaptAdamPreprint`, `DAdaptAdanIP`, and `DAdaptLion`.
- `DAdaptAdam` is now new. The old `DAdaptAdam` is available with `DAdaptAdamPreprint`.
- Simply specifying `DAdaptation` will use `DAdaptAdamPreprint` (same behavior as before).
- You need to install D-Adaptation v3.0. After activating venv, please do `pip install -U dadaptation`.
- See PR and D-Adaptation documentation for details.
- The state is saved even when the `--save_state` option is not specified in `fine_tune.py` and `train_db.py`. [PR #521](https://github.com/kohya-ss/sd-scripts/pull/521) Thanks to akshaal!
- Cannot load LoRA without `alpha`. [PR #527](https://github.com/kohya-ss/sd-scripts/pull/527) Thanks to Manjiz!
- Minor changes to console output during sample generation. [PR #515](https://github.com/kohya-ss/sd-scripts/pull/515) Thanks to yanhuifair!
- The generation script now uses xformers for VAE as well.
- Fixed an issue where an error would occur if the encoding of the prompt file was different from the default. [PR #510](https://github.com/kohya-ss/sd-scripts/pull/510) Thanks to sdbds!
- Added [English translation of documents](https://github.com/darkstorm2150/sd-scripts#links-to-usage-documentation) by darkstorm2150. Thank you very much!
- The prompt for sample generation during training can now be specified in `.toml` or `.json`. [PR #504](https://github.com/kohya-ss/sd-scripts/pull/504) Thanks to Linaqruf!
- For details on prompt description, please see the PR.
- Added an option `--dim_from_weights` to `train_network.py` to automatically determine the dim(rank) from the weight file. [PR #491](https://github.com/kohya-ss/sd-scripts/pull/491) Thanks to AI-Casanova!
- It is useful in combination with `resize_lora.py`. Please see the PR for details.
- Fixed a bug where the noise resolution was incorrect with Multires noise. [PR #489](https://github.com/kohya-ss/sd-scripts/pull/489) Thanks to sdbds!
- Please see the PR for details.
- The image generation scripts can now use img2img and highres fix at the same time.
- Fixed a bug where the hint image of ControlNet was incorrectly BGR instead of RGB in the image generation scripts.
- Added a feature to the image generation scripts to use the memory-efficient VAE.
- If you specify a number with the `--vae_slices` option, the memory-efficient VAE will be used. The maximum output size will be larger, but it will be slower. Please specify a value of about `16` or `32`.
- The implementation of the VAE is in `library/slicing_vae.py`.
- The documentation has been moved to the `docs` folder. If you have links, please change them.
- Removed `gradio` from `requirements.txt`.
- DAdaptAdaGrad, DAdaptAdan, and DAdaptSGD are now supported by DAdaptation. [PR#455](https://github.com/kohya-ss/sd-scripts/pull/455) Thanks to sdbds!
- DAdaptation needs to be installed. Also, depending on the optimizer, DAdaptation may need to be updated. Please update with `pip install --upgrade dadaptation`.
- Added support for pre-calculation of LoRA weights in image generation scripts. Specify `--network_pre_calc`.
- The prompt option `--am` is available. Also, it is disabled when Regional LoRA is used.
- Added Adaptive noise scale to each training script. Specify a number with `--adaptive_noise_scale` to enable it.
- __Experimental option. It may be removed or changed in the future.__
- This is an original implementation that automatically adjusts the value of the noise offset according to the absolute value of the mean of each channel of the latents. It is expected that appropriate noise offsets will be set for bright and dark images, respectively.
- Specify it together with `--noise_offset`.
- The actual value of the noise offset is calculated as `noise_offset + abs(mean(latents, dim=(2,3))) * adaptive_noise_scale`. Since the latent is close to a normal distribution, it may be a good idea to specify a value of about 1/10 to the same as the noise offset.
- Negative values can also be specified, in which case the noise offset will be clipped to 0 or more.
# set top parameter requires_grad = True for gradient checkpointing works
t_enc.text_model.embeddings.requires_grad_(True)
iftrain_text_encoder:
t_enc.text_model.embeddings.requires_grad_(True)
# set top parameter requires_grad = True for gradient checkpointing works
ifnottrain_text_encoder:# train U-Net only
unet.parameters().__next__().requires_grad_(True)
else:
unet.eval()
fort_encintext_encoders:
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
青龍聖者@bdsqlsz
GitHub