Auraflow Open-source Model - Achieve high-quality text-to-image conversion effects for free

Auraflow

Developed by ControlNetLoRA

A ControlNet PEFT LoRA model based on terminusresearch/auraflow-v0.3 for high-quality text-to-image conversion

Image Generation Open Source License:Apache-2.0 #ControlNet Fine-tuning #Text-to-Image Generation #LoRA Adapter

Downloads 142

Release Time : 6/17/2025

Model Overview

This is a LoRA adapter model based on the ControlNet architecture, mainly used for text-to-image generation tasks and capable of generating high-quality images in various scenarios.

Model Features

Efficient Fine-tuning

Adopt LoRA technology for parameter-efficient fine-tuning, significantly reducing the demand for training resources

High-Quality Image Generation

Capable of generating high-quality images at a resolution of 1024x1024

Low-Resource Inference

Supports int8 quantization to reduce the video memory requirement during inference

Model Capabilities

Text-to-Image Generation

High-Resolution Image Generation

Prompt-Based Image Synthesis

Use Cases

Creative Content Generation

Artistic Creation

Generate artworks based on text descriptions

High-quality images at a resolution of 1024x1024

Concept Design

Quickly generate product concept maps

Visual presentations that match the text descriptions

🚀 auraflow-controlnet-lora-test

This is a ControlNet PEFT LoRA derived from terminusresearch/auraflow-v0.3, which can be used for text - to - image and image - to - image tasks.

Information Table

Property	Details
Model Type	ControlNet PEFT LoRA
Base Model	terminusresearch/auraflow-v0.3
Pipeline Tag	text-to-image
Inference	true

🚀 Quick Start

The main validation prompt used during training was:

A photo-realistic image of a cat

✨ Features

Validation settings

CFG: 4.0
CFG Rescale: 0.0
Steps: 16
Sampler: FlowMatchEulerDiscreteScheduler
Seed: 42
Resolution: 1024x1024

Note: The validation settings are not necessarily the same as the training settings.

You can find some example images in the following gallery:

The text encoder was not trained. You may reuse the base model text encoder for inference.

Training settings

Training epochs: 15
Training steps: 450
Learning rate: 0.0001
- Learning rate schedule: constant
- Warmup steps: 500
Max grad value: 2.0
Effective batch size: 1
- Micro - batch size: 1
- Gradient accumulation steps: 1
- Number of GPUs: 1
Gradient checkpointing: True
Prediction type: flow_matching (extra parameters=['shift=3.0', 'controlnet_enabled'])
Optimizer: adamw_bf16
Trainable parameter precision: Pure BF16
Base model precision: int8 - torchao
Caption dropout probability: 0.0%
LoRA Rank: 64
LoRA Alpha: 64.0
LoRA Dropout: 0.1
LoRA initialisation style: default

Datasets

antelope-data-256

Repeats: 0
Total number of images: 29
Total number of aspect buckets: 1
Resolution: 0.065536 megapixels
Cropped: True
Crop style: center
Crop aspect: square
Used for regularisation data: No

💻 Usage Examples

Basic Usage

import torch
from diffusers import DiffusionPipeline

model_id = 'terminusresearch/auraflow-v0.3'
adapter_id = 'bghira/auraflow-controlnet-lora-test'
pipeline = DiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.bfloat16) # loading directly in bf16
pipeline.load_lora_weights(adapter_id)

prompt = "A photo-realistic image of a cat"
negative_prompt = 'ugly, cropped, blurry, low-quality, mediocre average'

## Optional: quantise the model to save on vram.
## Note: The model was quantised during training, and so it is recommended to do the same during inference time.
from optimum.quanto import quantize, freeze, qint8
quantize(pipeline.transformer, weights=qint8)
freeze(pipeline.transformer)
    
pipeline.to('cuda' if torch.cuda.is_available() else 'mps' if torch.backends.mps.is_available() else 'cpu') # the pipeline is already in its target precision level
model_output = pipeline(
    prompt=prompt,
    negative_prompt=negative_prompt,
    num_inference_steps=16,
    generator=torch.Generator(device='cuda' if torch.cuda.is_available() else 'mps' if torch.backends.mps.is_available() else 'cpu').manual_seed(42),
    width=1024,
    height=1024,
    guidance_scale=4.0,
).images[0]

model_output.save("output.png", format="PNG")

📄 License

This project is licensed under the Apache-2.0 license.

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご