Text-to-image finetuning - elephantmipt/test_tuned_sd_15

This pipeline was finetuned from stable-diffusion-v1-5/stable-diffusion-v1-5 on the None dataset. Below are some example images generated with the finetuned pipeline using the following prompts: ['IMAGE_ TYPE Cocktail Photography GENRE Coktail Shooting Lowlight EMOTION I want to drink it SCENE A beautiful and refreshing glass of a drink called lychee spritz , decorated set against a dreamy background lowlight, fitting to the image ACTORS None LOCATION TYPE Studio CAMERA MODEL Nikon D850 CAMERA LENSE 60mm f 2. 8 Macro SPECIAL EFFECTS Dreamy bokeh TIME_ OF_ DAY Studio lighting INTERACTION None ' 'Gandalf, Saruman, Radagast. Blue Wizards perform a captivating magic ritual intense focus, vibrant colors swirl like airborne gas. Mystical pentagram unites them. ' 'wide shot, desert, wall, nature, fuchsia pink, brick red, ochre yellow, pale pink, chipotle orange ' 'disney pixar style character, dodge challenger srt hellcat illustration drifting under the ocean, cartoon, super detail, no text, 8k, render 3d, wide view vision ' 'wide shoot of a typical farm in rural surroundings, near a clear water lake, beautiful flowers blooming , forest, saplings, moss, beautiful, epic lighting, ultrasharp, nikon 12mm f15 ' 'dramtic sky backgraund ' 'underwater lake, dusk, scarry, blue green bright shining, deep water, nessi, lake ness' 'Darkside Anakin Skywalker played by young Hayden Christensen with sith eyes, and a red lightsaber, hyperrealistic, cinematic, professional photo lighting, intricately detailed, cinematic lighting, 8k, ultra detailed, ultra realistic, photorealistic, camera Leica m11 quality with 30mm lens ']:

Pipeline usage

You can use the pipeline like so:

from diffusers import DiffusionPipeline
import torch

pipeline = DiffusionPipeline.from_pretrained("elephantmipt/test_tuned_sd_15", torch_dtype=torch.float16)
prompt = "IMAGE_ TYPE Cocktail Photography  GENRE Coktail Shooting Lowlight  EMOTION I want to drink it  SCENE A beautiful and refreshing glass of a drink called lychee spritz , decorated set against a dreamy background lowlight, fitting to the image  ACTORS None  LOCATION TYPE Studio  CAMERA MODEL Nikon D850  CAMERA LENSE 60mm f 2. 8 Macro  SPECIAL EFFECTS Dreamy bokeh  TIME_ OF_ DAY Studio lighting  INTERACTION None "
image = pipeline(prompt).images[0]
image.save("my_image.png")

Training info

These are the key hyperparameters used during training:

• Epochs: 14
• Learning rate: 8e-05
• Batch size: 20
• Gradient accumulation steps: 1
• Image resolution: 512
• Mixed-precision: bf16

More information on all the CLI arguments and the environment are available on your wandb run page.

Intended uses & limitations

How to use

# TODO: add an example code snippet for running this diffusion pipeline

Limitations and bias

[TODO: provide examples of latent issues and potential remediations]

Training details

[TODO: describe the data used to train the model]