# markos-diomataris/wandr architecture

> Architecture diagram of the GitHub repository markos-diomataris/wandr, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-02-03). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/markos-diomataris/wandr
- Repository: https://github.com/markos-diomataris/wandr

## Overview

##### Project type and purpose
- This is a research‐oriented Python machine‐learning pipeline (WANDR) for intention‐guided human motion generation.
- It provides data ingestion (SMPL-X body model), configuration via Hydra, model definition (CVAE, motion prior), training/evaluation loops, and rendering of generated motion.

##### File structure
- conf/ (Hydra config hierarchy): dataset, dataloader, model, optimizer, scheduler, logger, rendering, evaluator, paths, train/evaluate settings, custom callbacks.
- dataset/, dataloader, data/: raw data definitions, parsing, feature extraction.
- models/: abstract base modules and concrete implementations (motion_model, CVAE).
- motion_filter/: filters applied to raw motion sequences.
- schedulers/, utils/: supporting code (custom LR schedulers, math transforms, SMPL utilities).
- evaluators/: motion quality metrics.
- rendering/: utilities for visualizing and exporting motion.
- scripts/: environment setup (SMPL-X extraction, venv).
- Top‐level entry points: train.py (training), evaluate.py (quantitative evaluation), demo.py (inference + rendering), calculate_stats.py.

##### Extracted insights from the README
- External dependency: SMPL-X body model must be downloaded separately and put under data/body_models.
- setup.sh extracts the model and creates a virtual environment.
- demo.py produces an output.mp4 of generated motion.
- The project uses Hydra for config management, PyTorch (implied by modules), and optionally Weights & Biases for logging.

##### Diagramming recommendations
– Blue: Data (SMPL-X, deps, Dataset)
– Green: Core compute (Model, Filter, Training Loop)
– Orange: Configuration & Logging (Hydra, WandB)
– Purple: Output & Visualization (Evaluation, Rendering)

## Diagram (Mermaid)

```mermaid
flowchart TD
    subgraph "Configuration"
        HydraConfig["Hydra Config"]:::config
        HydraCallbacks["Hydra Callbacks"]:::config
    end

    subgraph "Data"
        SMPLX["SMPL-X Body Model"]:::data
        Deps["Precomputed Dependencies"]:::data
        Dataset["Dataset & DataLoader"]:::data
    end

    subgraph "Processing"
        Filter["Motion Filter"]:::compute
    end

    subgraph "Core Compute"
        BaseModule["Base Module"]:::compute
        BasePrior["Base Motion Prior"]:::compute
        MotionModel["Motion Model"]:::compute
        CVAE["CVAE / MotionPrior MLP"]:::compute
    end

    subgraph "Training"
        TrainLoop["Training Loop & Orchestration"]:::compute
        OptSched["Optimizers & Schedulers"]:::compute
    end

    subgraph "Output"
        Eval["Evaluation Module"]:::output
        Rendering["Rendering Engine"]:::output
    end

    Logging["Logging & Callbacks"]:::config
    Setup["Environment Setup Script"]:::config
    Utils["Utility Functions"]:::compute

    %% Configuration injection
    HydraConfig -->|"configures"| Dataset
    HydraConfig -->|"configures"| Filter
    HydraConfig -->|"configures"| BaseModule
    HydraConfig -->|"configures"| BasePrior
    HydraConfig -->|"configures"| MotionModel
    HydraConfig -->|"configures"| CVAE
    HydraConfig -->|"configures"| OptSched
    HydraConfig -->|"configures"| TrainLoop
    HydraConfig -->|"configures"| Eval
    HydraConfig -->|"configures"| Rendering
    HydraConfig -->|"configures"| Logging

    %% Data pipeline
    SMPLX -->|"raw body model"| Dataset
    Deps -->|"precomputed deps"| Dataset
    Dataset -->|"raw sequences"| Filter
    Filter -->|"filtered sequences"| TrainLoop

    %% Compute flow
    TrainLoop -->|"uses module"| BaseModule
    TrainLoop -->|"uses prior"| BasePrior
    TrainLoop -->|"trains"| MotionModel
    TrainLoop -->|"predicts"| CVAE
    TrainLoop -->|"updates"| OptSched

    %% Outputs
    TrainLoop -->|"checkpoints & metrics"| Logging
    TrainLoop -->|"trained model"| Eval
    TrainLoop -->|"trained model"| Rendering

    %% Utilities and setup
    Setup -->|"sets up env"| HydraConfig
    Setup -->|"provides scripts"| Dataset
    Utils -->|"supports"| Filter
    Utils -->|"supports"| BaseModule
    Utils -->|"supports"| MotionModel

    %% Click Events
    click HydraConfig "https://github.com/markos-diomataris/wandr/tree/main/conf/"
    click HydraCallbacks "https://github.com/markos-diomataris/wandr/blob/main/my_hydra_callbacks/mycallbacks.py"
    click SMPLX "https://github.com/markos-diomataris/wandr/tree/main/data/body_models/"
    click Deps "https://github.com/markos-diomataris/wandr/tree/main/deps/"
    click Dataset "https://github.com/markos-diomataris/wandr/tree/main/dataset/"
    click Filter "https://github.com/markos-diomataris/wandr/blob/main/motion_filter/amass_filter.py"
    click BaseModule "https://github.com/markos-diomataris/wandr/blob/main/models/base_module.py"
    click BasePrior "https://github.com/markos-diomataris/wandr/blob/main/models/base_motion_prior.py"
    click MotionModel "https://github.com/markos-diomataris/wandr/blob/main/models/motion_model.py"
    click CVAE "https://github.com/markos-diomataris/wandr/blob/main/models/motion_prior_mlp_cvae_goal.py"
    click TrainLoop "https://github.com/markos-diomataris/wandr/blob/main/train.py"
    click OptSched "https://github.com/markos-diomataris/wandr/blob/main/schedulers/my_lr_scheduler.py"
    click Eval "https://github.com/markos-diomataris/wandr/blob/main/evaluators/motion_evaluator.py"
    click Rendering "https://github.com/markos-diomataris/wandr/blob/main/rendering/render_utils.py"
    click Logging "https://github.com/markos-diomataris/wandr/tree/main/conf/logger/"
    click Setup "https://github.com/markos-diomataris/wandr/blob/main/scripts/setup.sh"
    click Utils "https://github.com/markos-diomataris/wandr/tree/main/utils/"

    %% Styles
    classDef data fill:#ADD8E6,stroke:#000
    classDef compute fill:#98FB98,stroke:#000
    classDef config fill:#FFA500,stroke:#000
    classDef output fill:#DDA0DD,stroke:#000
```
