Text-to-image finetuning - rcannizzaro/sd-dsprites-counterfactual-with-vae-loss-vae-unfrozen

This pipeline was finetuned from runwayml/stable-diffusion-v1-5 on the osazuwa/dsprite-counterfactual dataset. Below are some example images generated with the finetuned pipeline using the following prompts: ['A very large square, with no rotation, at the center horizontally and vertically.', 'A very large ellipse, with no rotation, at the center horizontally and vertically.', 'A very large heart shape, with no rotation, at the center horizontally and vertically.']:

Pipeline usage

You can use the pipeline like so:

from diffusers import DiffusionPipeline
import torch

pipeline = DiffusionPipeline.from_pretrained("rcannizzaro/sd-dsprites-counterfactual-with-vae-loss-vae-unfrozen", torch_dtype=torch.float16)
prompt = "A very large square, with no rotation, at the center horizontally and vertically."
image = pipeline(prompt).images[0]
image.save("my_image.png")

Training info

These are the key hyperparameters used during training:

• Epochs: 3
• Learning rate: 1e-05
• Batch size: 100
• Gradient accumulation steps: 4
• Image resolution: 64
• Mixed-precision: fp16

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]