Command_Line_Interface_CLI_.md

graph LR
    Command_Line_Interface_CLI_["Command-Line Interface (CLI)"]
    Training_CLI_Entrypoint["Training CLI Entrypoint"]
    Rendering_CLI_Entrypoint["Rendering CLI Entrypoint"]
    Mesh_Extraction_CLI_Entrypoint["Mesh Extraction CLI Entrypoint"]
    Training_Orchestrator["Training Orchestrator"]
    Training_Launcher["Training Launcher"]
    Core_Training_Loop["Core Training Loop"]
    Core_Rendering_Logic["Core Rendering Logic"]
    Command_Line_Interface_CLI_ -- "contains" --> Training_CLI_Entrypoint
    Command_Line_Interface_CLI_ -- "contains" --> Rendering_CLI_Entrypoint
    Command_Line_Interface_CLI_ -- "contains" --> Mesh_Extraction_CLI_Entrypoint
    Training_CLI_Entrypoint -- "initiates" --> Training_Orchestrator
    Training_Orchestrator -- "orchestrates" --> Training_Launcher
    Training_Launcher -- "calls" --> Core_Training_Loop
    Rendering_CLI_Entrypoint -- "delegates to" --> Core_Rendering_Logic
    click Command_Line_Interface_CLI_ href "https://github.com/CodeBoarding/GeneratedOnBoardings/blob/main/sdfstudio/Command_Line_Interface_CLI_.md" "Details"

Loading

Details

The Command-Line Interface (CLI) subsystem serves as the primary user interface for sdfstudio, orchestrating various operations such as model training, scene rendering, and mesh extraction. It acts as the gateway to the core ML pipelines, providing distinct entry points for each major functionality.

Command-Line Interface (CLI) [Expand]

The overarching user interface, providing the main entry points for all sdfstudio operations. It parses high-level commands and dispatches them to specific functional modules.

Related Classes/Methods:

• scripts.train
• scripts.render
• scripts.extract_mesh

Training CLI Entrypoint

The specific entry point for initiating model training via the train command. It handles initial argument parsing and setup before delegating to the main training orchestration logic.

Related Classes/Methods:

• scripts.train.entrypoint:246-255

Rendering CLI Entrypoint

The specific entry point for rendering scenes or trajectories via the render command. It processes rendering-specific arguments and orchestrates the rendering pipeline.

Related Classes/Methods:

• scripts.render.main:126-159

Mesh Extraction CLI Entrypoint

The specific entry point for extracting 3D meshes from trained models via the extract_mesh command. It manages the parameters for mesh extraction and initiates the process.

Related Classes/Methods:

• scripts.extract_mesh.main:62-133

Training Orchestrator

Orchestrates the overall training workflow, setting up the environment, loading configurations, and preparing the training process. It acts as a high-level manager for the training pipeline.

Related Classes/Methods:

• scripts.train.main:220-243

Training Launcher

Responsible for launching the actual training process, often involving distributed training setup or process management. It bridges the orchestration layer with the core training loop.

Related Classes/Methods:

• scripts.train.launch:149-217

Core Training Loop

Contains the iterative core training logic, including forward and backward passes, optimization steps, and metric logging. This is where the model learning occurs.

Related Classes/Methods:

• scripts.train.train_loop:78-90

Core Rendering Logic

Handles the detailed process of rendering a trajectory or scene into a video format. It interacts with the rendering engine and model to generate visual outputs.

Related Classes/Methods:

• scripts.render._render_trajectory_video:38-100

FAQ