keras-dreambooth/ignatius

Model description

The Ignatius Farray dreambooth model would be a sleek and modern diffusion model designed to transport users into a world of absurdity and hilarity. I cannot promise that all the images would be adorned with bright, eye-catching colors and images that reflect Ignatius' unique sense of style and humor.

Images generated by model

Intended uses & limitations

You can use to create images based on Ignatius and put him in different situations. Try not to use for bad purpose and use the "commedia" on it.

Training and evaluation data

To train this model, this was the training notebook and the trainig dataset was this one

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

Hyperparameters	Value
inner_optimizer.class_name	Custom>RMSprop
inner_optimizer.config.name	RMSprop
inner_optimizer.config.weight_decay	None
inner_optimizer.config.clipnorm	None
inner_optimizer.config.global_clipnorm	None
inner_optimizer.config.clipvalue	None
inner_optimizer.config.use_ema	False
inner_optimizer.config.ema_momentum	0.99
inner_optimizer.config.ema_overwrite_frequency	100
inner_optimizer.config.jit_compile	True
inner_optimizer.config.is_legacy_optimizer	False
inner_optimizer.config.learning_rate	0.0010000000474974513
inner_optimizer.config.rho	0.9
inner_optimizer.config.momentum	0.0
inner_optimizer.config.epsilon	1e-07
inner_optimizer.config.centered	False
dynamic	True
initial_scale	32768.0
dynamic_growth_steps	2000
training_precision	mixed_float16

Model Plot

View Model Plot

Usage

The instance token used is "ignatius". A prompt example is as follows "a photo of ignatius on a car"

from huggingface_hub import from_pretrained_keras
import keras_cv

sd_dreambooth_model = keras_cv.models.StableDiffusion(
    img_width=resolution, img_height=resolution, jit_compile=True, 
)
loaded_diffusion_model = from_pretrained_keras("keras-dreambooth/ignatius")
sd_dreambooth_model._diffusion_model = loaded_diffusion_model

prompt = f"ignatius on the moon"

#generated_img = sd_dreambooth_model.text_to_image(
generated_img = dreambooth_model.text_to_image(
    prompt,
    batch_size=4,
    num_steps=150,
    unconditional_guidance_scale=15,
)