AI Toolkit Docker Image

Docker image for training FLUX.1 and other models, based on AI Toolkit.

Quick Start

Shared Drive: Mount lora-training-shared at /app/workspace

Captioning Setup

# Navigate to /app/workspace/captions

# Make the setup script executable:
chmod +x setup_caption_env.sh

# Run the setup:
./setup_caption_env.sh

After setup:

• Place images in /app/workspace/captions/input/

• Run captioning process from captions/ directory

FLUX.1 Training

Requirements

• GPU with ≥24GB VRAM

⚠️ Important: Don't interrupt during checkpoint saves