my-pet-cat

🐱 My Pet Cat | Stable Diffusion Model

A fine-tuned Stable Diffusion model tailored to generate adorable images of a specific pet cat. > Turn your text prompts into fluffy reality.

---

🖼️ Gallery

Sample generations created with this model:

---

🚀 Quick Start

Get started generating images in seconds using the diffusers library.

1. Install Dependencies

pip install diffusers transformers torch

2. Python Usage

import torch
from diffusers import StableDiffusionPipeline

model_id = "irkky/my-pet-cat"

# Load the pipeline
pipe = StableDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16)
pipe = pipe.to("cuda")

# Run inference
# Note: Try using 'xzg cat' or specific tokens if applicable to the training data
prompt = "A cute xzg cat playing with a ball of yarn, cinematic lighting, 4k"
image = pipe(prompt).images[0]

# Save your creation
image.save("my_cat_creation.png")

---

🛠️ Model Details

---

📂 Repository Structure

The model weights are organized for use with the Diffusers library:

• unet/: The U-Net architecture (denoising).
• text_encoder/: The CLIP text encoder.
• vae/: Variational Autoencoder for decoding latents.
• tokenizer/: Tokenizer files.
• scheduler/: Scheduler configuration.
• xzg.ckpt: Original checkpoint file (2.61 GB).

---

⚠️ Limitations & Bias

• Specific Subject: This model is overfitted to a specific cat identity; general cat prompts may look different than base Stable Diffusion.
• Resolution: Best performance at 512x512 resolution.
• Content: The model inherits the safety and bias limitations of the original Stable Diffusion v1.5.

---