--- license: creativeml-openrail-m tags: - pytorch - diffusers - stable-diffusion - text-to-image - diffusion-models-class - dreambooth-hackathon - animal widget: - text: a photo of a ðŁĴŁ jellyfish in the snow - text: a photo of a ðŁĴŁ jellyfish next to a dog - text: a photo of a ðŁĴŁ jellyfish on top of a mountain --- # Iridescent Jellyfish **Iridescent Jellyfish** is a Dreambooth model for the `iridescent` jellyfish concept (represented by the `ðŁĴŁ` identifier). It applies to the *animal* theme. It is fine-tuned from `runwayml/stable-diffusion-v1-5` checkpoint on a small dataset of jellyfish images. It can be used by modifying the `instance_prompt`: **a photo of a ðŁĴŁ jellyfish in the snow** This model was created as part of the DreamBooth Hackathon 🔥. Visit the [organisation page](https://huggingface.co/dreambooth-hackathon) for instructions on how to take part! #### Fine-Tuning Details - Number of training images: 17 - Learning rate: 2e-06 - Training steps: 800 - Guidance Scale: 7 - Inference Steps: 50 #### Output Examples
a oil painting of a ðŁĴŁ jellyfish a photo of a ðŁĴŁ jellyfish next to a dog a photo of a ðŁĴŁ jellyfish in the snow
a photo of a ðŁĴŁ jellyfish on top of a mountain a photo of a ðŁĴŁ jellyfish in the sky a photo of a ðŁĴŁ jellyfish
a photo of a ðŁĴŁ jellyfish skydiving a photo of a ðŁĴŁ jellyfish sutfing on a surfboard a photo of a choclate ðŁĴŁ jellyfish
a photo of a ðŁĴŁ jellyfish shooting fireworks in the sky a photo of a ðŁĴŁ jellyfish on rollerblades a photo of a ðŁĴŁ jellyfish in a beer bottle
a colorful sketch of a ðŁĴŁ jellyfish a photo of a ðŁĴŁ jellyfish in the jungle a mystic ðŁĴŁ jellyfish, trending on artstation
## Usage ```python from diffusers import StableDiffusionPipeline import torch device = torch.device('cuda:0' if torch.cuda.is_available() else 'cpu') pipeline = StableDiffusionPipeline.from_pretrained('simonschoe/iridescent-jellyfish').to(device) prompt = "a photo of a ðŁĴŁ jellyfish in the snow" image = pipeline( prompt, num_inference_steps=50, guidance_scale=7, num_images_per_prompt=1 ).images[0] image ```