--- license: other license_name: bespoke-lora-trained-license license_link: >- https://multimodal.art/civitai-licenses?allowNoCredit=True&allowCommercialUse=Rent&allowDerivatives=True&allowDifferentLicense=False tags: - text-to-image - stable-diffusion - lora - diffusers - template:sd-lora - disney - pixar - dreamworks - style - cartoon - sdxl style lora base_model: stabilityai/stable-diffusion-xl-base-1.0 instance_prompt: DreamDisPix style widget: - text: 'Shrek playing with olaf in a frozen field DreamDisPix style ' output: url: 4857923.jpeg - text: 'The girl from the incredibles shooting a laser gun DreamDisPix style ' output: url: 4857954.jpeg - text: 'The girl with a pearl earring DreamDisPix style ' output: url: 4857994.jpeg - text: 'Snoop Dogg DreamDisPix style ' output: url: 4857995.jpeg - text: 'A mighty dragon at the dentist DreamDisPix style ' output: url: 4858005.jpeg - text: 'A cute dragon playing with Eevee DreamDisPix style ' output: url: 4857996.jpeg - text: 'Zombie Cinderella has rissen from the dead DreamDisPix style ' output: url: 4858004.jpeg - text: 'Santa riding a might dragon DreamDisPix style ' output: url: 4858026.jpeg - text: 'Pocahontas riding a Zebra DreamDisPix style ' output: url: 4858031.jpeg - text: 'American gothic DreamDisPix style ' output: url: 4858028.jpeg datasets: - Norod78/DreamDisPix-blip2-captions --- # Dream Dis Pix XL ([CivitAI](https://civitai.com/models/242711)) ## Model description

A (kinda failed) attempt to train an SDXL LoRA model upon on a mixed style dataset of images from Dreamworks, Disney and Pixar. I have uploaded this Blip2-Captioned Dataset to Huggingface.

I usually prefer to train in LoRA ranks 4-16 because so far it always seemed to be enough and provided a reasonably small checkpoint but in this case, I only started getting results when I went for Rank-32, even tried 24 and it was not enough.

My conclusion is that it should have probably been trained as several different lower ranked LoRAs each for the specific visual sub-style and it would have probably looked way better for each.

## Trigger words You should use `DreamDisPix style` to trigger the image generation. ## Download model Weights for this model are available in Safetensors format. [Download](/Norod78/dream-dis-pix-xl/tree/main) them in the Files & versions tab. ## Use it with the [🧨 diffusers library](https://github.com/huggingface/diffusers) ```py from diffusers import AutoPipelineForText2Image import torch pipeline = AutoPipelineForText2Image.from_pretrained('stabilityai/stable-diffusion-xl-base-1.0', torch_dtype=torch.float16).to('cuda') pipeline.load_lora_weights('Norod78/dream-dis-pix-xl', weight_name='SDXL-DreamDisPix-Lora-r32.safetensors') image = pipeline('American gothic DreamDisPix style ').images[0] ``` For more details, including weighting, merging and fusing LoRAs, check the [documentation on loading LoRAs in diffusers](https://huggingface.co/docs/diffusers/main/en/using-diffusers/loading_adapters)