sd-wikiart-v2 is a stable diffusion model that has been fine-tuned on the
wikiart dataset to generate artistic images in different style and genres.
The model originally used for fine-tuning is
Stable Diffusion V1-4, which is a latent image diffusion model trained on
LAION2B-en.
The current model has been fine-tuned with a learning rate of 1e-05 for 1 epoch on 81K text-image pairs from wikiart dataset. Only the attention layers of the model are fine-tuned. This is done to avoid catastrophic forgetting, the model can generate artistic images given specific prompts but still retains most of its previous knowledge.
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:
This model can be used for entertainment purposes and as a generative art assistant.
1import torch
2from diffusers import StableDiffusionPipeline
3
4model_id = "valhalla/sd-wikiart-v2"
5device = "cuda"
6
7pipe = StableDiffusionPipeline.from_pretrained(
8 model_id,
9 torch_dtype=torch.float16,
10)
11pipe = pipe.to(device)
12
13prompt = "a painting of eiffel tower in the style of surrealism"
14with torch.autocast("cuda"):
15 image = pipe(prompt, guidance_scale=7.5).images[0]
16
17image.save("eiffel_impressionism.png")
