This is a fine-tuned Stable Diffusion model for generating composite sketches of human faces from text descriptions. It allows reflecting both specific facial features as well as abstract impressions mentioned in the text input.
usage
python
1prompt ="남성. 40대. 얼굴은 사각형이고 턱은 각진형(사각)이다. 광대가 나왔다. 짧은 머리이고 왼쪽가르마를 탔다. 눈썹이 진하고 작은 눈에 코가 크고 인중은 길다. 입이 작다. 팔자주름이 있다. 관리하지 않은 헤어나 피부에서 상당한 피로감이 보인다. 째려보는 듯 째진눈에서 날카로운 통찰을 할 것 같은 느낌이 있고, 형사또는 건축의 총괄을 할 것 같은 차분하면서 치밀한 성격의 이미지이다."2seed =35463generator = torch.Generator("cuda").manual_seed(seed)
examples
Image 1-6 demonstrate variations on a core description (Image 0) to evaluate the model's ability to reflect subtle changes within the prompt.
Images 1 to 6 were generated by partially varying the text prompt used for image 0.
Image 0
"남성. 40대. 얼굴은 사각형이고 턱은 각진형(사각)이다. 광대가 나왔다. 짧은 머리이고 왼쪽가르마를 탔다. 눈썹이 진하고 작은 눈에 코가 크고 인중은 길다. 입이 작다. 팔자주름이 있다. 관리하지 않은 헤어나 피부에서 상당한 피로감이 보인다. 째려보는 듯 째진눈에서 날카로운 통찰을 할 것 같은 느낌이 있고, 형사또는 건축의 총괄을 할 것 같은 차분하면서 치밀한 성격의 이미지이다."
("Male. 40s. Has a square face and a sharp jawline. High cheekbones. Short hair and a left-side parting. Has thick eyebrows, small eyes, a large nose, and a long philtrum. Small mouth. Has smile lines. He looks quite tired, with ungroomed hair and skin. He has a sharp gaze and seems to be very perceptive. He seems to be a calm and meticulous person, perhaps a detective or a construction manager.")
img_0
Image 1
“40대” → “20대” ("40s" → "20s")
img_1
Image 2
“남성” → ”여성” ("male" → "female")
img_2
Image 3
얼굴 “사각형”→”계란형”, 턱 “각진형”→”둥근형” ("square face and a sharp jawline" → "oval shaped face and a round jawline")
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
Note: This section is taken from the DALLE-MINI model card, but applies in the same way to Stable Diffusion v1.
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 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/ 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),
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 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 CLIPTextModelafter 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.
