|
--- |
|
tags: |
|
- pytorch |
|
- diffusers |
|
- stable-diffusion |
|
- text-to-image |
|
- diffusion-models-class |
|
- dreambooth-hackathon |
|
--- |
|
|
|
# DreamBooth model for the groot concept traine on the LinoyTsaban/dreambooth-hackathon-images-groot dataset. |
|
|
|
This is a Stable Diffusion model fine-tuned on the groot concept with DreamBooth. It can be used by modifying the `instance_prompt`: groot figurine |
|
|
|
This model was created as part of the DreamBooth Hackathon 🔥. Visit the [organisation page](https://huggingface.co/dreambooth-hackathon) for instructions on how to take part! |
|
|
|
params used: |
|
scheduler- DDIMScheduler |
|
resolution=512 |
|
learning_rate=2e-06 |
|
max_train_steps= 500 |
|
train_batch_size=1 |
|
gradient_accumulation_steps=2 |
|
max_grad_norm=1.0 |
|
gradient_checkpointing=True |
|
use_8bit_adam=True |
|
seed=14071995, |
|
sample_batch_size=2 |
|
|
|
|
|
## Usage |
|
|
|
```python |
|
from diffusers import StableDiffusionPipeline |
|
|
|
pipeline = StableDiffusionPipeline.from_pretrained('LinoyTsaban/dreambooth-groot') |
|
name_of_concept = "groot" |
|
type_of_thing = "figurine" |
|
|
|
prompt = f"a photo of groot figurine in the Louvre museum" |
|
|
|
# Tune the guidance to control how closely the generations follow the prompt. |
|
# Values between 7-11 usually work best |
|
guidance_scale = 8 # |
|
image = pipeline(prompt, guidance_scale=guidance_scale).images[0] |
|
image |
|
``` |
|
|
