README.md

---
license: openrail++
base_model: runwayml/stable-diffusion-v1-5
tags:
- stable-diffusion
- stable-diffusion-diffusers
- text-to-image
- diffusers
- lora
- cartoon
- illustration
datasets:
- juliaturc/captioned-cartoons
---

# cartoon-lora-aggressive-medium

Fine-tuned LoRA model for generating colorful cartoon-style illustrations.

## Model Details

- **Base Model:** runwayml/stable-diffusion-v1-5
- **Dataset:** juliaturc/captioned-cartoons
- **Training Framework:** Diffusers + LoRA
- **Task:** Text-to-Image Generation (Cartoon Style)

## Training Configuration

```json
{
  "learning_rate": 0.0001,
  "max_train_steps": 1500,
  "rank": 48,
  "lr_scheduler": "cosine"
}
```

## Usage

### Install dependencies

```bash
pip install diffusers transformers torch
```

### Load and generate

```python
import torch
from diffusers import StableDiffusionPipeline

# Load base model
model_id = "runwayml/stable-diffusion-v1-5"
pipe = StableDiffusionPipeline.from_pretrained(
    model_id,
    torch_dtype=torch.float16
)
pipe = pipe.to("cuda")

# Load LoRA weights
pipe.load_lora_weights("yzhdvbQ/cartoon-lora-aggressive-medium")

# Generate image
prompt = "Two people walking in a park with dogs, colorful cartoon style, woman in pink shirt, man in green vest, fall scenery"
image = pipe(prompt, num_inference_steps=30).images[0]
image.save("output.png")
```

## Example Prompts

1. `Two people walking in a park with dogs, colorful cartoon style, woman in pink shirt, man in green vest, fall scenery`
2. `A young woman sitting on a bench reading a book in a city park, wearing a yellow sweater and jeans, autumn trees, colorful cartoon illustration`
3. `A happy family having a picnic in a meadow, parents and two children, red checkered blanket, colorful cartoon style`


## Training Details

This model was fine-tuned on a cartoon illustration dataset with detailed captions describing:
- Characters and their appearance
- Clothing and accessories
- Actions and poses
- Environment and scenery
- Composition and style

The training used Low-Rank Adaptation (LoRA) to efficiently fine-tune Stable Diffusion while keeping the base model frozen.

## Limitations

- Works best with detailed, descriptive prompts
- Optimized for cartoon/illustration style
- May not perform well on photorealistic requests
- Based on Stable Diffusion 1.5 architecture

## Citation

```bibtex
@misc{cartoon-lora-ai54,
  author = {AI54 Course},
  title = {Cartoon LoRA for Stable Diffusion},
  year = {2026},
  publisher = {HuggingFace},
  url = {https://huggingface.co/yzhdvbQ/cartoon-lora-aggressive-medium}
}
```

## License

This model inherits the license from Stable Diffusion (OpenRAIL++)