Text-to-image finetuning - zachary-shah/mri-bruno-sd-v2_base-512-bs128-zerotermsnr

This pipeline was finetuned from yurman/mri_full_512_v2_base on the stanford dataset for brain image generation. Below are some example images generated with the finetuned pipeline:

Pipeline usage

You can use the pipeline like so:

from diffusers import DiffusionPipeline
import torch

pipeline = DiffusionPipeline.from_pretrained("zachary-shah/mri-bruno-sd-v2_base-512-bs128-zerotermsnr", torch_dtype=torch.float16)
prompt = "An empty, flat black image with a MRI brain axial scan in the center"
image = pipeline(prompt).images[0]
image.save("my_image.png")

Training info

These are the key hyperparameters used during training:

• Epochs: 173
• Learning rate: 5e-05
• embeds rate: 1e-05
• Batch size: 8
• Classifier free guidance: 1
• VAE scaling: Same as in the original model
• Input perturbation: 0
• Noise offset: 0
• Gradient accumulation steps: 4
• Image resolution: 512
• Mixed-precision: None

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