How fallenshock/flowedit works
Main Components Identification
a) Input Processing Layer
• Image input handling
• Text prompt processing
• Configuration management (YAML files)
b) Core Processing Layer
• Flow Model Integration
• Text-to-Image Pipeline
• Editing Pipeline
c) Output Generation Layer
• Image Generation/Modification
• Result Storage
[Input Layer]
• Image Input (example_images/)
• Source Prompt (edits.yaml)
• Target Prompt (edits.yaml)
• Configuration (SD3_exp.yaml/FLUX_exp.yaml)
[Processing Layer]
• Flow Model Controller
• FLUX Implementation
• Stable Diffusion 3 Implementation
• Text Processing Pipeline
• Source Prompt Encoding
• Target Prompt Encoding
• Image Processing Pipeline
• Inversion-Free Processing
• Flow Model Transformation
[Output Layer]
• Modified Image Output
Important Details
• Model Integration Points:
• Connections to external models (FLUX, SD3)
• Where configuration parameters (n_max, n_min) are applied
• Data Flow Indicators:
• Image path flow
• Prompt processing flow
• Model processing flow