Latent Diffusion Models (LDM) for super-resolution

Paper: High-Resolution Image Synthesis with Latent Diffusion Models

Abstract:

By decomposing the image formation process into a sequential application of denoising autoencoders, diffusion models (DMs) achieve state-of-the-art synthesis results on image data and beyond. Additionally, their formulation allows for a guiding mechanism to control the image generation process without retraining. However, since these models typically operate directly in pixel space, optimization of powerful DMs often consumes hundreds of GPU days and inference is expensive due to sequential evaluations. To enable DM training on limited computational resources while retaining their quality and flexibility, we apply them in the latent space of powerful pretrained autoencoders. In contrast to previous work, training diffusion models on such a representation allows for the first time to reach a near-optimal point between complexity reduction and detail preservation, greatly boosting visual fidelity. By introducing cross-attention layers into the model architecture, we turn diffusion models into powerful and flexible generators for general conditioning inputs such as text or bounding boxes and high-resolution synthesis becomes possible in a convolutional manner. Our latent diffusion models (LDMs) achieve a new state of the art for image inpainting and highly competitive performance on various tasks, including unconditional image generation, semantic scene synthesis, and super-resolution, while significantly reducing computational requirements compared to pixel-based DMs.

Authors

Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, Björn Ommer

Checkpoints: CompVis/ldm-super-resolution-4x-openimages

Introduction: This model designed to upscale images by a factor of 4, enhancing resolution and detail.

Usage

Sample code

import requests
from PIL import Image
from io import BytesIO
from clarifai.client.model import Model

inference_params = dict(
  num_inference_steps = 50, # A higher value improves image quality but increases processing time. Recommended range: 50–100.
  eta=1
)

url = "https://samples.clarifai.com/car-dashboard-steering-wheel.jpg"
image = requests.get(url).content

# Model Predict
model = Model("https://clarifai.com/comp-vis/super-resolution/models/ldm-super-resolution-4x-openimages-2")
prediction = model.predict_by_bytes(image, input_type="image", inference_params=inference_params)
image_bytes = prediction.outputs[0].data.image.base64

# Do something else with the output
output_image = Image.open(BytesIO(image_bytes))
output_image.save("tmp.jpg")

Inference parameters:

• num_inference_steps (int): The number of denoising steps. More denoising steps usually lead to a higher quality image at the expense of slower inference. Defaults to 50.
• eta (float): Corresponds to parameter eta (η) from the DDIM paper. Only applies to the DDIMScheduler, and is ignored in other schedulers. Defaults to 1.0.