# nvlabs/vt-refine architecture

> Architecture diagram of the GitHub repository nvlabs/vt-refine, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-10-02). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/nvlabs/vt-refine
- Repository: https://github.com/nvlabs/vt-refine

## Overview

VT-Refine releases workflows for tactile-policy pre-training, evaluation, and simulation fine-tuning for bimanual assembly. The README identifies `dppo/script/run.py` as the entry point and documents the stages and data-fetch command. The tree names the `easysim` and `easysim-envs` submodules, but no source excerpts show their runtime interfaces, so their links are limited to the README’s documented simulation fine-tuning relationship.

Main entry points: [`dppo`](https://github.com/nvlabs/vt-refine/tree/main/dppo) (Fine-tuning), [`easysim`](https://github.com/nvlabs/vt-refine/tree/main/easysim) (Simulation runtime), [`data/fetch_vt-refine_data.sh`](https://github.com/nvlabs/vt-refine/blob/main/data/fetch_vt-refine_data.sh) (Data fetch), [`easysim-envs`](https://github.com/nvlabs/vt-refine/tree/main/easysim-envs) (Simulation tasks)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_workflow["Policy workflows"]
  node_data_fetch["Data fetch"]
  node_runner["Workflow runner"]
  node_pretrain["Pre-training"]
  node_pretrain_eval["Pretrain evaluation"]
  node_finetune["Fine-tuning"]
  node_finetune_eval["Fine-tune evaluation"]
end

subgraph group_simulation["Simulation"]
  node_sim["Simulation runtime"]
  node_sim_envs["Simulation tasks"]
end

node_researcher(("Researcher"))

node_researcher -->|"fetches data"| node_data_fetch
node_researcher -->|"runs workflows"| node_runner
node_runner -->|"runs pre-training"| node_pretrain
node_runner -->|"runs evaluation"| node_pretrain_eval
node_runner -->|"runs fine-tuning"| node_finetune
node_runner -->|"runs evaluation"| node_finetune_eval
node_finetune -.->|"uses simulation"| node_sim
node_sim -.->|"uses tasks"| node_sim_envs

click node_data_fetch "https://github.com/nvlabs/vt-refine/blob/main/data/fetch_vt-refine_data.sh"
click node_pretrain "https://github.com/nvlabs/vt-refine/tree/main/dppo"
click node_pretrain_eval "https://github.com/nvlabs/vt-refine/tree/main/dppo"
click node_finetune "https://github.com/nvlabs/vt-refine/tree/main/dppo"
click node_finetune_eval "https://github.com/nvlabs/vt-refine/tree/main/dppo"
click node_sim "https://github.com/nvlabs/vt-refine/tree/main/easysim"
click node_sim_envs "https://github.com/nvlabs/vt-refine/tree/main/easysim-envs"

classDef toneNeutral fill:#f8fafc,stroke:#334155,stroke-width:1.5px,color:#0f172a
classDef toneBlue fill:#dbeafe,stroke:#2563eb,stroke-width:1.5px,color:#172554
classDef toneAmber fill:#fef3c7,stroke:#d97706,stroke-width:1.5px,color:#78350f
classDef toneMint fill:#dcfce7,stroke:#16a34a,stroke-width:1.5px,color:#14532d
classDef toneRose fill:#ffe4e6,stroke:#e11d48,stroke-width:1.5px,color:#881337
classDef toneIndigo fill:#e0e7ff,stroke:#4f46e5,stroke-width:1.5px,color:#312e81
classDef toneTeal fill:#ccfbf1,stroke:#0f766e,stroke-width:1.5px,color:#134e4a
class node_data_fetch,node_runner,node_pretrain,node_pretrain_eval,node_finetune,node_finetune_eval toneBlue
class node_sim,node_sim_envs toneAmber
class node_researcher toneIndigo
```

## Components

### Policy workflows

- [Data fetch](https://github.com/nvlabs/vt-refine/blob/main/data/fetch_vt-refine_data.sh): `data/fetch_vt-refine_data.sh`
- **Workflow runner**
- [Pre-training](https://github.com/nvlabs/vt-refine/tree/main/dppo): `dppo`
- [Pretrain evaluation](https://github.com/nvlabs/vt-refine/tree/main/dppo): `dppo`
- [Fine-tuning](https://github.com/nvlabs/vt-refine/tree/main/dppo): `dppo`
- [Fine-tune evaluation](https://github.com/nvlabs/vt-refine/tree/main/dppo): `dppo`

### Simulation

- [Simulation runtime](https://github.com/nvlabs/vt-refine/tree/main/easysim): `easysim`
- [Simulation tasks](https://github.com/nvlabs/vt-refine/tree/main/easysim-envs): `easysim-envs`

### Other components

- **Researcher**

## Connections

- Researcher → Data fetch: fetches data (evidence: `README.md`)
- Researcher → Workflow runner: runs workflows (evidence: `README.md`)
- Workflow runner → Pre-training: runs pre-training (evidence: `README.md`)
- Workflow runner → Pretrain evaluation: runs evaluation (evidence: `README.md`)
- Workflow runner → Fine-tuning: runs fine-tuning (evidence: `README.md`)
- Workflow runner → Fine-tune evaluation: runs evaluation (evidence: `README.md`)
- Fine-tuning → Simulation runtime: uses simulation (evidence: `README.md`)
- Simulation runtime → Simulation tasks: uses tasks
