# invoke-ai/invoke-training architecture

> Architecture diagram of the GitHub repository invoke-ai/invoke-training, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2025-09-10). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/invoke-ai/invoke-training
- Repository: https://github.com/invoke-ai/invoke-training

## Overview

##### Project type and purpose
- This is an open-source Python library (invoke-training) for configuring and launching custom Stable Diffusion training workflows (fine-tuning, LoRA, textual inversion, etc.), usable via both a CLI and a GUI.
- The README highlights core use cases: stable diffusion (SD) training modes, SDXL training modes, and integration with InvokeAI.

##### File structure
- Top-level directories:
- src/invoke_training – the main codebase
- src/invoke_training/_shared – shared utilities (accelerator, data, checkpoints, optimizer, flux integrations, stable_diffusion helpers)
- src/invoke_training/config – pydantic-style pipeline and data/optimizer config definitions
- src/invoke_training/pipelines – orchestration of various training modes (stable_diffusion, stable_diffusion_xl, flux, experimental)
- src/invoke_training/model_merge – tools to extract and merge LoRA weights
- src/invoke_training/scripts – convenience wrappers and experiments
- src/invoke_training/ui – GRadio-based GUI (app, config groups, blocks, pages)
- src/invoke_training/sample_configs – YAML examples for different GPUs and modes
- tests – unit tests for every module
- CI/CD: .github/workflows (test, deploy), pre-commit, mkdocs.yml for documentation.
- Docs live under docs/ (architecture, getting started, guides, API reference).

##### README for additional insights
- Architecture is config-driven: pipelines accept YAML configs, build data loaders, optimizers, model components, run training loops.
- Dual interface: CLI (“invoke-train”) and GUI (“invoke-train-ui”) expose the same underlying pipeline engine.
- Training progress streamed to TensorBoard. Models saved as LoRA or full checkpoints, compatible with InvokeAI.

##### Detail
- Layer: interface, config, orchestration, compute, persistence, monitoring, extensions.

## Diagram (Mermaid)

```mermaid
flowchart TD
    %% User Interfaces
    subgraph "User Interfaces"
        CLI["CLI Entrypoint"]:::interface
        GUI["GUI Entrypoint"]:::interface
        GRadioApp["GRadio App (app.py)"]:::interface
    end

    %% Configuration Subsystem
    subgraph "Configuration Subsystem"
        ConfigModels["Pydantic Config Models"]:::config
        SampleConfigs["YAML Sample Configs"]:::config
    end

    %% Pipeline Orchestrator
    subgraph "Pipeline Orchestrator"
        PipelineOrch["Pipeline Dispatcher & Plugins"]:::orchestrator
    end

    %% Core Services
    subgraph "Core Services"
        DataSvc["Data Service"]:::service
        ModelSvc["Model Service"]:::service
        AccelSvc["Accelerator Service"]:::service
        OptimizerSvc["Optimizer Service"]:::service
        CheckpointSvc["Checkpoint Service"]:::service
    end

    %% Monitoring & Persistence
    subgraph "Monitoring & Persistence"
        TensorBoard["TensorBoard Integration"]:::monitoring
        FS["File System (Checkpoints, Logs)"]:::persistence
    end

    %% Extensions
    subgraph "Extensions"
        ModelMerge["Model-Merge Subsystem"]:::extension
    end

    %% Data Flows
    CLI -->|uses config| ConfigModels
    GUI -->|uses config| ConfigModels
    CLI -->|loads samples| SampleConfigs
    GUI -->|loads samples| SampleConfigs

    ConfigModels -->|instantiates| PipelineOrch
    SampleConfigs -->|feeds| ConfigModels

    PipelineOrch -->|calls| DataSvc
    PipelineOrch -->|calls| ModelSvc
    PipelineOrch -->|calls| AccelSvc
    PipelineOrch -->|calls| OptimizerSvc
    PipelineOrch -->|calls| CheckpointSvc
    PipelineOrch -->|logs metrics| TensorBoard
    CheckpointSvc -->|writes to| FS
    TensorBoard -->|writes events| FS

    CLI -->|can trigger| ModelMerge
    GUI -->|can trigger| ModelMerge
    ModelMerge -->|reads from| FS
    ModelMerge -->|writes to| FS

    PipelineOrch -->|integrates callbacks| TensorBoard

    %% Click Events
    click CLI "https://github.com/invoke-ai/invoke-training/blob/main/src/invoke_training/scripts/invoke_train.py"
    click GUI "https://github.com/invoke-ai/invoke-training/blob/main/src/invoke_training/scripts/invoke_train_ui.py"
    click GRadioApp "https://github.com/invoke-ai/invoke-training/blob/main/src/invoke_training/ui/app.py"
    click ConfigModels "https://github.com/invoke-ai/invoke-training/tree/main/src/invoke_training/config/"
    click SampleConfigs "https://github.com/invoke-ai/invoke-training/tree/main/src/invoke_training/sample_configs/"
    click PipelineOrch "https://github.com/invoke-ai/invoke-training/tree/main/src/invoke_training/pipelines/"
    click DataSvc "https://github.com/invoke-ai/invoke-training/tree/main/src/invoke_training/_shared/data/"
    click ModelSvc "https://github.com/invoke-ai/invoke-training/tree/main/src/invoke_training/_shared/stable_diffusion/"
    click AccelSvc "https://github.com/invoke-ai/invoke-training/blob/main/src/invoke_training/_shared/accelerator/accelerator_utils.py"
    click OptimizerSvc "https://github.com/invoke-ai/invoke-training/blob/main/src/invoke_training/_shared/optimizer/optimizer_utils.py"
    click CheckpointSvc "https://github.com/invoke-ai/invoke-training/tree/main/src/invoke_training/_shared/checkpoints/"
    click TensorBoard "https://github.com/invoke-ai/invoke-training/blob/main/src/invoke_training/pipelines/callbacks.py"
    click ModelMerge "https://github.com/invoke-ai/invoke-training/tree/main/src/invoke_training/model_merge/"

    %% Styles
    classDef interface fill:#AEDFF7,stroke:#0288D1,stroke-width:1px
    classDef config fill:#FFF2AE,stroke:#DFAF00,stroke-width:1px
    classDef orchestrator fill:#D1F7C4,stroke:#34A853,stroke-width:1px
    classDef service fill:#F7D6D6,stroke:#EA4335,stroke-width:1px
    classDef monitoring fill:#E8D9F7,stroke:#8E24AA,stroke-width:1px
    classDef persistence fill:#F0F0F0,stroke:#888888,stroke-width:1px
    classDef extension fill:#FFE0B2,stroke:#FB8C00,stroke-width:1px
```
