# internrobotics/internnav architecture

> Architecture diagram of the GitHub repository internrobotics/internnav, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2025-09-29). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/internrobotics/internnav
- Repository: https://github.com/internrobotics/internnav

## Overview

##### Project Type and Purpose
- InternNav is an all-in-one open-source toolbox for embodied navigation. It provides a modular framework—built on PyTorch, Habitat, and Isaac Sim—for training, evaluating, and deploying vision-language and visual navigation agents.

##### Top-Level Components (from file structure & README)
a. CLI / Scripting Layer
- scripts/train/train.py + configs → Training entrypoint
- scripts/eval/eval.py + configs → Evaluation entrypoint
- challenge/health_check.sh, start_eval_iros.sh → Challenge evaluation drivers
- scripts/realworld/http_internvla_{client,server}.py → Real-world HTTP service
b. Core Library (internnav/)
- configs/         → Hydra-style configuration modules (agent, trainer, model, evaluator)
- dataset/         → Data loaders & adapters (LMDB, custom dataset formats)
- env/             → Environment interfaces (VLN-multi, VLN-PE)
- habitat_extensions/ → Extensions to Habitat for custom measures, maps, fonts
- agent/           → High-level agent orchestrators (cma_agent, rdp_agent, seq2seq_agent)
- model/           →
– basemodel/   → Policy implementations for different algorithm backbones (CMA, RDP, Seq2Seq, diffusion-based)
– encoder/     → Vision & language encoders (CLIP, BERT, ResNet, depth models, diffusion policy encoders)
– utils/       → Shared utilities across models
- trainer/         → Trainer classes (cma_trainer, navdp_trainer, rdp_trainer, etc.)
- evaluator/       → Evaluator classes & utilities (vln_multi_evaluator, vln_pe_evaluator, metrics collection)
- utils/           → Logging, checkpoint I/O, tensorboard, distributed utils, visualization
c. Data Conversion & Miscellaneous
- scripts/dataset_converters/ → Convert between VLN-CE, LeRobot formats
- internnav/dist.py → Distributed training helpers
d. Tests
- tests/           → Unit & end-to-end functional tests

##### External Dependencies & Services
- PyTorch: model implementation, training loops
- Habitat & Isaac Sim: simulators for virtual environments
- LMDB: high-throughput dataset storage
- Gradio (via backend script): web UI for evaluation demo
- HTTP server/client: real-world deployment interface

## Diagram (Mermaid)

```mermaid
flowchart TD
    %% CLI Layer
    subgraph "CLI Layer"
        TrainEntr["train.py"]:::cli
        TrainCfg["train/configs"]:::cli
        EvalEntr["eval.py"]:::cli
        EvalCfg["eval/configs"]:::cli
        Health["health_check.sh"]:::cli
        StartEval["start_eval_iros.sh"]:::cli
        HTTPClient["http_internvla_client.py"]:::cli
        HTTPServer["http_internvla_server.py"]:::cli
    end

    %% Core Library
    subgraph "Core Library"
        subgraph "Config Manager"
            Config["internnav/configs"]:::core
        end
        subgraph "Trainer Module"
            Trainer["internnav/trainer"]:::core
        end
        subgraph "Evaluator Module"
            Evaluator["internnav/evaluator"]:::core
            HabitatExt["internnav/habitat_extensions"]:::core
        end
        subgraph "Agent Orchestrator"
            Agent["internnav/agent"]:::core
        end
        subgraph "Model Zoo"
            BaseModel["internnav/model/basemodel"]:::core
            Encoder["internnav/model/encoder"]:::core
            ModelUtils["internnav/model/utils"]:::core
        end
        subgraph "Dataset Layer"
            Dataset["internnav/dataset"]:::core
        end
        subgraph "Environment Adapters"
            Env["internnav/env"]:::core
        end
        subgraph "Utilities & IPC"
            Utils["internnav/utils"]:::core
            Dist["internnav/dist.py"]:::core
            IPC["shared_memory"]:::data
        end
        subgraph "Data Conversion & Misc"
            Converters["scripts/dataset_converters"]:::cli
            LeRobot["lerobot_as_lmdb.py"]:::core
        end
        subgraph "Tests"
            Tests["tests"]:::core
        end
    end

    %% External Services
    subgraph "External Services"
        PyTorch["PyTorch"]:::external
        Habitat["Habitat"]:::external
        Isaac["Isaac Sim"]:::external
        LMDB["LMDB"]:::data
        Gradio["Gradio"]:::external
    end

    %% Training Pipeline
    TrainEntr --> Config
    Config --> Trainer
    Trainer --> Dataset
    Trainer --> Env
    Dataset --> Agent
    Env --> Agent
    Agent --> BaseModel
    Agent --> Encoder
    Agent --> ModelUtils
    Trainer --> PyTorch
    Trainer --> Dist
    Env --> LMDB
    Env --> Habitat
    Env --> Isaac

    %% Evaluation Pipeline
    EvalEntr --> Config
    Config --> Evaluator
    Evaluator --> Env
    Env --> Evaluator
    Evaluator --> Agent
    Evaluator --> HabitatExt
    Evaluator --> Utils
    Evaluator --> Gradio
    Evaluator --> Tests

    %% Real-World HTTP Branch
    HTTPServer --> Agent
    Agent --> HTTPClient

    %% Click Events
    click TrainEntr "https://github.com/internrobotics/internnav/blob/main/scripts/train/train.py"
    click TrainCfg "https://github.com/internrobotics/internnav/tree/main/scripts/train/configs/"
    click EvalEntr "https://github.com/internrobotics/internnav/blob/main/scripts/eval/eval.py"
    click EvalCfg "https://github.com/internrobotics/internnav/tree/main/scripts/eval/configs/"
    click Health "https://github.com/internrobotics/internnav/blob/main/challenge/health_check.sh"
    click StartEval "https://github.com/internrobotics/internnav/blob/main/challenge/start_eval_iros.sh"
    click HTTPClient "https://github.com/internrobotics/internnav/blob/main/scripts/realworld/http_internvla_client.py"
    click HTTPServer "https://github.com/internrobotics/internnav/blob/main/scripts/realworld/http_internvla_server.py"
    click Config "https://github.com/internrobotics/internnav/tree/main/internnav/configs/"
    click Trainer "https://github.com/internrobotics/internnav/tree/main/internnav/trainer/"
    click Evaluator "https://github.com/internrobotics/internnav/tree/main/internnav/evaluator/"
    click HabitatExt "https://github.com/internrobotics/internnav/tree/main/internnav/habitat_extensions/"
    click Agent "https://github.com/internrobotics/internnav/tree/main/internnav/agent/"
    click BaseModel "https://github.com/internrobotics/internnav/tree/main/internnav/model/basemodel/"
    click Encoder "https://github.com/internrobotics/internnav/tree/main/internnav/model/encoder/"
    click ModelUtils "https://github.com/internrobotics/internnav/tree/main/internnav/model/utils/"
    click Dataset "https://github.com/internrobotics/internnav/tree/main/internnav/dataset/"
    click Env "https://github.com/internrobotics/internnav/tree/main/internnav/env/"
    click Utils "https://github.com/internrobotics/internnav/tree/main/internnav/utils/"
    click Dist "https://github.com/internrobotics/internnav/blob/main/internnav/dist.py"
    click IPC "https://github.com/internrobotics/internnav/tree/main/internnav/model/encoder/diffusion_policy/shared_memory/"
    click Converters "https://github.com/internrobotics/internnav/tree/main/scripts/dataset_converters/"
    click LeRobot "https://github.com/internrobotics/internnav/blob/main/internnav/utils/lerobot_as_lmdb.py"
    click Tests "https://github.com/internrobotics/internnav/tree/main/tests/"

    %% Styles
    classDef core fill:#D6EAF8,stroke:#1B4F72,color:#1B2631;
    classDef cli fill:#F9E79F,stroke:#B9770E,color:#7E5109;
    classDef external fill:#ABEBC6,stroke:#1D8348,color:#145A32;
    classDef data fill:#E8DAEF,stroke:#6C3483,color:#4A235A
```
