A Complete Guide to Image Generation in Python with Hugging Face Diffusers

📘 Introduction

Generating images with AI has never been easier. Hugging Face’s Diffusers library provides a user-friendly way to create stunning visuals using pre-trained diffusion models like Stable Diffusion. In this guide, we’ll walk through the entire process step by step.

🧠 What are Hugging Face Diffusers?

Diffusers are models that iteratively refine random noise into a coherent image based on a given text prompt. Hugging Face’s diffusers library makes it easy to access these models and generate high-quality images with just a few lines of code.

✅ Prerequisites

Before you begin, make sure you have:

🐍☑️ Installed Python
🌐☑️ Created and activated a virtual environment (venv)

📦1️⃣ Install Libraries

Install the following Python packages using pip:

pip install diffusers
pip install transformers
pip install torch torchvision torchaudio
pip install accelerate

📥2️⃣ Import Libraries

Start by importing the required Python modules:

import torch
import transformers
from diffusers import DiffusionPipeline

💻3️⃣ Set Device

Choose whether to run the model on GPU or CPU. On Macs with M1/M2 chips, PyTorch can use Metal Performance Shaders (MPS) for GPU acceleration:

device = "mps" if torch.backends.mps.is_available() else "cpu"

If MPS is available, the model runs on the Mac GPU for faster image generation; otherwise, it defaults to CPU. This ensures optimal speed and efficiency across different hardware.

⚙️4️⃣ Load the Pre-trained Model

Load a ready-to-use Stable Diffusion model from Hugging Face. Using a pre-trained model means we don’t need to train one from scratch—it already knows how to generate high-quality images from text prompts.

pipeline = DiffusionPipeline.from_pretrained(
    "runwayml/stable-diffusion-v1-5",
    torch_dtype=torch.float16
).to(device)

You should get the following output: