|
--- |
|
license: cc-by-sa-4.0 |
|
datasets: |
|
- foldl/rumeme-desc |
|
language: |
|
- ru |
|
library_name: diffusers |
|
pipeline_tag: text-to-image |
|
--- |
|
|
|
# ruMeme Description model |
|
|
|
|
|
## Example |
|
|
|
*Use `Diffusers` >=0.8.0, do not support lower versions.* |
|
|
|
```python |
|
from diffusers import StableDiffusionPipeline |
|
import torch |
|
|
|
model_path = "foldl/sd-rumeme-desc" |
|
pipe = StableDiffusionPipeline.from_pretrained(model_path, torch_dtype=torch.float16) |
|
pipe.to("cuda") |
|
|
|
image = pipe(prompt="кот").images[0] |
|
image.save("cat".jpg) |
|
``` |
|
|
|
## Training |
|
|
|
### Training Procedure |
|
|
|
Model was trained on 1 P100 GPU for 10k steps. |
|
|
|
Base model - https://huggingface.co/OFA-Sys/small-stable-diffusion-v0 |
|
|
|
Training notebook here - https://www.kaggle.com/code/nukeee/meme-diffusion |
|
|
|
### Training Data |
|
- https://huggingface.co/datasets/foldl/rumeme-desc |
|
- Same dataset at [Kaggle](https://www.kaggle.com/datasets/nukeee/rumemes-descriptions) |
