ERNIE-Image Diffusers API Guide: Python Code from Zero to Production

HuggingFace Diffusers fully supports ERNIE-Image — learn to call ERNIE-Image with Python for image generation.

---

Installation

pip install diffusers transformers accelerate torch

pip install git+https://github.com/huggingface/diffusers.git

---

Basic Usage

Single Image Generation

from diffusers import ERNIEImagePipeline
import torch

# Load model
pipe = ERNIEImagePipeline.from_pretrained(
    "baidu/ERNIE-Image",
    torch_dtype=torch.float16
)
pipe = pipe.to("cuda")

# Generate
image = pipe(
    prompt="a cat sitting on a windowsill, city view at night",
    negative_prompt="",
    width=1024,
    height=1024,
    num_inference_steps=28,
    guidance_scale=7.0,
    generator=torch.Generator("cuda").manual_seed(42)
).images[0]

image.save("output.png")

---

Batch Generation

from diffusers import ERNIEImagePipeline
import torch

pipe = ERNIEImagePipeline.from_pretrained(
    "baidu/ERNIE-Image",
    torch_dtype=torch.float16
).to("cuda")

prompts = [
    "a cat on a table",
    "a dog in a garden",
    "a bird in a tree",
    "a fish in a pond"
]

images = pipe(
    prompt=prompts,
    negative_prompt=[""] * len(prompts),
    width=1024,
    height=1024,
    num_inference_steps=28,
    guidance_scale=7.0
).images

for i, img in enumerate(images):
    img.save(f"output_{i}.png")

---

Turbo Mode

from diffusers import ERNIEImageTurboPipeline

pipe = ERNIEImageTurboPipeline.from_pretrained(
    "baidu/ERNIE-Image-Turbo",
    torch_dtype=torch.float16
).to("cuda")

# Turbo: 8 steps
image = pipe(
    prompt="a sunset over the ocean, cinematic lighting",
    width=1024,
    height=1024,
    num_inference_steps=8,
    guidance_scale=1.0
).images[0]

image.save("turbo_output.png")

---

Inpainting

from diffusers import ERNIEImageInpaintPipeline
from PIL import Image

pipe = ERNIEImageInpaintPipeline.from_pretrained(
    "baidu/ERNIE-Image",
    torch_dtype=torch.float16
).to("cuda")

init_image = Image.open("input.png").convert("RGB")
mask_image = Image.open("mask.png").convert("L")

image = pipe(
    prompt="a beautiful garden with flowers",
    image=init_image,
    mask_image=mask_image,
    width=1024,
    height=1024,
    num_inference_steps=28,
    guidance_scale=7.0
).images[0]

image.save("inpaint_output.png")

---

PE Module

from diffusers import ERNIEImagePipeline

pipe = ERNIEImagePipeline.from_pretrained(
    "baidu/ERNIE-Image",
    torch_dtype=torch.float16,
    enable_pe=True
).to("cuda")

image = pipe(
    prompt="cat on window",
    width=1024,
    height=1024,
    num_inference_steps=28,
    guidance_scale=7.0
).images[0]

image.save("pe_output.png")

---

API Service Deployment

from fastapi import FastAPI
from pydantic import BaseModel
from diffusers import ERNIEImagePipeline
import torch

app = FastAPI()
pipe = ERNIEImagePipeline.from_pretrained(
    "baidu/ERNIE-Image",
    torch_dtype=torch.float16
).to("cuda")

class GenerateRequest(BaseModel):
    prompt: str
    width: int = 1024
    height: int = 1024
    steps: int = 28

@app.post("/generate")
async def generate(req: GenerateRequest):
    image = pipe(
        prompt=req.prompt,
        width=req.width,
        height=req.height,
        num_inference_steps=req.steps,
        guidance_scale=7.0
    ).images[0]
    
    import io
    buf = io.BytesIO()
    image.save(buf, format="PNG")
    buf.seek(0)
    
    from fastapi.responses import Response
    return Response(buf.read(), media_type="image/png")

---

Summary

Diffusers ERNIE-Image key points:

1. Install: diffusers + transformers + accelerate
2. Basic: Pipeline.from_pretrained → pipe() → save
3. Batch: Pass prompt list
4. Turbo: 8 steps, no high CFG needed
5. Inpainting: Mask-guided local regeneration
6. API: FastAPI + Diffusers = full service

Diffusers provides a clean API, making ERNIE-Image integration simple.

---

Based on HuggingFace Diffusers + ERNIE-Image 8B model.