--- license: creativeml-openrail-m tags: - stable-diffusion - stable-diffusion-diffusers - text-to-image - core-ml pipeline_tag: text-to-image extra_gated_prompt: >- This model is open access and available to all, with a CreativeML OpenRAIL-M license further specifying rights and usage. The CreativeML OpenRAIL License specifies: 1. You can't use the model to deliberately produce nor share illegal or harmful outputs or content 2. The authors claim no rights on the outputs you generate, you are free to use them and are accountable for their use which must not go against the provisions set in the license 3. You may re-distribute the weights and use the model commercially and/or as a service. If you do, please be aware you have to include the same use restrictions as the ones in the license and share a copy of the CreativeML OpenRAIL-M to all your users (please read the license entirely and carefully) Please read the full license carefully here: https://huggingface.co/spaces/CompVis/stable-diffusion-license extra_gated_heading: Please read the LICENSE to access this model --- # Palettized Core ML Weights for BK-SDMs For deployment on iOS 17 or macOS 14, this model card introduces **palettized Core ML weights** of BK-SDM-{[Base-2M](https://huggingface.co/nota-ai/bk-sdm-base-2m), [Small-2M](https://huggingface.co/nota-ai/bk-sdm-small-2m) and [Tiny-2M](https://huggingface.co/nota-ai/bk-sdm-tiny-2m)}. These weights were generated using [Apple’s repository](https://github.com/apple/ml-stable-diffusion) which has [ASCL](https://github.com/apple/ml-stable-diffusion/blob/main/LICENSE.md). - Block-removed Knowledge-distilled Stable Diffusion Model (BK-SDM) is an architecturally compressed SDM for efficient text-to-image synthesis. This model is bulit with (i) removing several residual and attention blocks from the U-Net of [Stable Diffusion v1.4]( https://huggingface.co/CompVis/stable-diffusion-v1-4) and (ii) distillation pretraining. Despite being trained with very limited resources, our compact model can imitate the original SDM by benefiting from transferred knowledge. - More information about BK-SDMs: [Paper](https://arxiv.org/abs/2305.15798), [GitHub](https://github.com/Nota-NetsPresso/BK-SDM), [Demo]( https://huggingface.co/spaces/nota-ai/compressed-stable-diffusion). A demo to use Core ML Stable Diffusion weights can be found [here](https://github.com/huggingface/swift-coreml-diffusers). ## Deployment Results | Base Model Name | Pipeline Size | Quantization Type | Attention Implementation | |----------------------|---------------|-----------------------------------------------------------------------------------|--------------------------| | [BK-SDM-Base-2M](https://huggingface.co/nota-ai/bk-sdm-base-2m) (Ours) | 1.48GB | [Palettized](https://coremltools.readme.io/docs/palettization-overview) | split_einsum_v2 | | [BK-SDM-Small-2M](https://huggingface.co/nota-ai/bk-sdm-small-2m) (Ours) | 1.44GB | [Palettized](https://coremltools.readme.io/docs/palettization-overview) | split_einsum_v2 | | [BK-SDM-Tiny-2M](https://huggingface.co/nota-ai/bk-sdm-tiny-2m) (Ours) | 1.43GB | [Palettized](https://coremltools.readme.io/docs/palettization-overview) | split_einsum_v2 | | [OFA-Sys' Small Stable Diffusion v0](https://huggingface.co/pcuenq/coreml-small-stable-diffusion-v0) | 3.28GB | None | split_einsum | | [Apple's Stable Diffusion v1.4, Palettized](https://huggingface.co/apple/coreml-stable-diffusion-1-4-palettized) | 1.57GB | [Palettized](https://coremltools.readme.io/docs/palettization-overview) | split_einsum_v2 | ## Compression Method ### U-Net Architecture Certain residual and attention blocks were eliminated from the U-Net of SDM-v1.4: - 1.04B-param [SDM-v1.4](https://huggingface.co/CompVis/stable-diffusion-v1-4) (0.86B-param U-Net): the original source model. - 0.76B-param [**BK-SDM-Base**](https://huggingface.co/nota-ai/bk-sdm-base) (0.58B-param U-Net): obtained with ① fewer blocks in outer stages. - 0.66B-param [**BK-SDM-Small**](https://huggingface.co/nota-ai/bk-sdm-small) (0.49B-param U-Net): obtained with ① and ② mid-stage removal. - 0.50B-param [**BK-SDM-Tiny**](https://huggingface.co/nota-ai/bk-sdm-tiny) (0.33B-param U-Net): obtained with ①, ②, and ③ further inner-stage removal.
U-Net architectures
### Distillation Pretraining The compact U-Net was trained to mimic the behavior of the original U-Net. We leveraged feature-level and output-level distillation, along with the denoising task loss.
KD-based pretraining

# Uses _Note: This section is taken from the [Stable Diffusion v1 model card]( https://huggingface.co/CompVis/stable-diffusion-v1-4) (which was based on the [DALLE-MINI model card](https://huggingface.co/dalle-mini/dalle-mini)) and applies in the same way to BK-SDMs_. ## Direct Use The model is intended for research purposes only. Possible research areas and tasks include - Safe deployment of models which have the potential to generate harmful content. - Probing and understanding the limitations and biases of generative models. - Generation of artworks and use in design and other artistic processes. - Applications in educational or creative tools. - Research on generative models. Excluded uses are described below. ### Misuse, Malicious Use, and Out-of-Scope Use The model should not be used to intentionally create or disseminate images that create hostile or alienating environments for people. This includes generating images that people would foreseeably find disturbing, distressing, or offensive; or content that propagates historical or current stereotypes. #### Out-of-Scope Use The model was not trained to be factual or true representations of people or events, and therefore using the model to generate such content is out-of-scope for the abilities of this model. #### Misuse and Malicious Use Using the model to generate content that is cruel to individuals is a misuse of this model. This includes, but is not limited to: - Generating demeaning, dehumanizing, or otherwise harmful representations of people or their environments, cultures, religions, etc. - Intentionally promoting or propagating discriminatory content or harmful stereotypes. - Impersonating individuals without their consent. - Sexual content without consent of the people who might see it. - Mis- and disinformation - Representations of egregious violence and gore - Sharing of copyrighted or licensed material in violation of its terms of use. - Sharing content that is an alteration of copyrighted or licensed material in violation of its terms of use. ## Limitations and Bias ### Limitations - The model does not achieve perfect photorealism - The model cannot render legible text - The model does not perform well on more difficult tasks which involve compositionality, such as rendering an image corresponding to “A red cube on top of a blue sphere” - Faces and people in general may not be generated properly. - The model was trained mainly with English captions and will not work as well in other languages. - The autoencoding part of the model is lossy - The model was trained on a large-scale dataset [LAION-5B](https://laion.ai/blog/laion-5b/) which contains adult material and is not fit for product use without additional safety mechanisms and considerations. - No additional measures were used to deduplicate the dataset. As a result, we observe some degree of memorization for images that are duplicated in the training data. The training data can be searched at [https://rom1504.github.io/clip-retrieval/](https://rom1504.github.io/clip-retrieval/) to possibly assist in the detection of memorized images. ### Bias While the capabilities of image generation models are impressive, they can also reinforce or exacerbate social biases. Stable Diffusion v1 was trained on subsets of [LAION-2B(en)](https://laion.ai/blog/laion-5b/), which consists of images that are primarily limited to English descriptions. Texts and images from communities and cultures that use other languages are likely to be insufficiently accounted for. This affects the overall output of the model, as white and western cultures are often set as the default. Further, the ability of the model to generate content with non-English prompts is significantly worse than with English-language prompts. ### Safety Module The intended use of this model is with the [Safety Checker](https://github.com/huggingface/diffusers/blob/main/src/diffusers/pipelines/stable_diffusion/safety_checker.py) in Diffusers. This checker works by checking model outputs against known hard-coded NSFW concepts. The concepts are intentionally hidden to reduce the likelihood of reverse-engineering this filter. Specifically, the checker compares the class probability of harmful concepts in the embedding space of the `CLIPTextModel` *after generation* of the images. The concepts are passed into the model with the generated image and compared to a hand-engineered weight for each NSFW concept. # Acknowledgments - We express our gratitude to [Microsoft for Startups Founders Hub](https://www.microsoft.com/en-us/startups) for generously providing the Azure credits used during pretraining. - We deeply appreciate the pioneering research on Latent/Stable Diffusion conducted by [CompVis](https://github.com/CompVis/latent-diffusion), [Runway](https://runwayml.com/), and [Stability AI](https://stability.ai/). - We also appreciate the effort from [Apple](https://www.apple.com) for the [Core ML Stable Diffusion](https://github.com/apple/ml-stable-diffusion) library. - Special thanks to the contributors to [LAION](https://laion.ai/), [Diffusers](https://github.com/huggingface/diffusers), and [Gradio](https://www.gradio.app/) for their valuable support. # Citation ```bibtex @article{kim2023architectural, title={On Architectural Compression of Text-to-Image Diffusion Models}, author={Kim, Bo-Kyeong and Song, Hyoung-Kyu and Castells, Thibault and Choi, Shinkook}, journal={arXiv preprint arXiv:2305.15798}, year={2023}, url={https://arxiv.org/abs/2305.15798} } ``` ```bibtex @article{Kim_2023_ICMLW, title={BK-SDM: Architecturally Compressed Stable Diffusion for Efficient Text-to-Image Generation}, author={Kim, Bo-Kyeong and Song, Hyoung-Kyu and Castells, Thibault and Choi, Shinkook}, journal={ICML Workshop on Efficient Systems for Foundation Models (ES-FoMo)}, year={2023}, url={https://openreview.net/forum?id=bOVydU0XKC} } ``` *This model card was written by Thibault Castells and is based on the [bk-sdm-base model card](https://huggingface.co/nota-ai/bk-sdm-base) and the [coreml-stable-diffusion-1-4-palettized model card](https://huggingface.co/apple/coreml-stable-diffusion-1-4-palettized).*