# julienexr/autoencoders-lab architecture

> Architecture diagram of the GitHub repository julienexr/autoencoders-lab, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-11-21). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/julienexr/autoencoders-lab
- Repository: https://github.com/julienexr/autoencoders-lab

## Overview

This repository is a PyTorch autoencoder implementation suite with CLI training and an interactive Streamlit demo. The training workflow selects a model and dataset, creates dataloaders, trains the chosen model, and saves checkpoints; the demo loads models for reconstruction or generation. The model families are distinct branches, with VQ-VAE extended by a Transformer prior. README and excerpts establish the main capabilities; unsampled training and app code limit some call-level relationships, so those links are documented or omitted rather than inferred.

Main entry points: [`app.py`](https://github.com/julienexr/autoencoders-lab/blob/main/app.py) (Interactive demo), [`main.py`](https://github.com/julienexr/autoencoders-lab/blob/main/main.py) (Training CLI), [`src/visualization.py`](https://github.com/julienexr/autoencoders-lab/blob/main/src/visualization.py) (Visual outputs), [`modules/vq_vae/training.py`](https://github.com/julienexr/autoencoders-lab/blob/main/modules/vq_vae/training.py) (VQ-VAE training)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_entry["User entry points"]
  node_cli["Training CLI<br/>[main.py]"]
  node_demo["Interactive demo<br/>[app.py]"]
end

subgraph group_data["Data and outputs"]
  node_datasets["Image datasets<br/>[data.py]"]
  node_checkpoints[("Model checkpoints")]
  node_visuals["Visual outputs<br/>[visualization.py]"]
end

subgraph group_models["Autoencoder models"]
  node_ae["Autoencoder<br/>[ae.py]"]
  node_vae["Variational AE<br/>[vae.py]"]
  node_cvae["Conditional VAE<br/>[cvae.py]"]
  node_vqvae["Vector-quantized AE<br/>[vq_vae.py]"]
  node_prior["Transformer prior"]
end

subgraph group_train["Training and visualization"]
  node_ae_train["AE training<br/>[training.py]"]
  node_vae_train["VAE training<br/>[training.py]"]
  node_cvae_train["CVAE training<br/>[training.py]"]
  node_vq_train["VQ-VAE training<br/>[training.py]"]
  node_inference["VQ prior inference<br/>[inference.py]"]
  node_loss["Perceptual loss<br/>[loss.py]"]
end

node_user(("User"))

node_user -->|"chooses training"| node_cli
node_user -->|"opens demo"| node_demo
node_cli -->|"loads data"| node_datasets
node_cli -->|"trains AE"| node_ae_train
node_cli -->|"trains VAE"| node_vae_train
node_cli -->|"saves weights"| node_checkpoints
node_ae_train -->|"renders outputs"| node_visuals
node_vae_train -->|"renders outputs"| node_visuals
node_vae_train -->|"computes perceptual loss"| node_loss
node_cvae_train -->|"renders outputs"| node_visuals
node_vq_train -->|"renders outputs"| node_visuals
node_vq_train -->|"encodes batches"| node_vqvae
node_vq_train -->|"trains prior"| node_prior
node_inference -->|"samples indices"| node_prior
node_inference -->|"decodes samples"| node_vqvae
node_demo -->|"loads examples"| node_datasets
node_demo -.->|"runs model"| node_ae
node_demo -.->|"runs model"| node_vae
node_demo -.->|"runs model"| node_cvae
node_demo -.->|"runs model"| node_vqvae
node_demo -.->|"uses prior"| node_prior
node_demo -.->|"loads weights"| node_checkpoints
node_datasets -.->|"downloads datasets"| node_user

click node_cli "https://github.com/julienexr/autoencoders-lab/blob/main/main.py"
click node_demo "https://github.com/julienexr/autoencoders-lab/blob/main/app.py"
click node_datasets "https://github.com/julienexr/autoencoders-lab/blob/main/src/data.py"
click node_checkpoints "https://github.com/julienexr/autoencoders-lab/tree/main/models"
click node_visuals "https://github.com/julienexr/autoencoders-lab/blob/main/src/visualization.py"
click node_ae "https://github.com/julienexr/autoencoders-lab/blob/main/modules/autoencoder/ae.py"
click node_vae "https://github.com/julienexr/autoencoders-lab/blob/main/modules/vae/vae.py"
click node_cvae "https://github.com/julienexr/autoencoders-lab/blob/main/modules/cvae/cvae.py"
click node_vqvae "https://github.com/julienexr/autoencoders-lab/blob/main/modules/vq_vae/vq_vae.py"
click node_prior "https://github.com/julienexr/autoencoders-lab/blob/main/modules/vq_vae/tranformer_prior.py"
click node_ae_train "https://github.com/julienexr/autoencoders-lab/blob/main/modules/autoencoder/training.py"
click node_vae_train "https://github.com/julienexr/autoencoders-lab/blob/main/modules/vae/training.py"
click node_cvae_train "https://github.com/julienexr/autoencoders-lab/blob/main/modules/cvae/training.py"
click node_vq_train "https://github.com/julienexr/autoencoders-lab/blob/main/modules/vq_vae/training.py"
click node_inference "https://github.com/julienexr/autoencoders-lab/blob/main/modules/vq_vae/inference.py"
click node_loss "https://github.com/julienexr/autoencoders-lab/blob/main/src/loss.py"

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_cli,node_demo,node_user toneBlue
class node_datasets,node_checkpoints,node_visuals toneAmber
class node_ae,node_vae,node_cvae,node_vqvae,node_prior toneMint
class node_ae_train,node_vae_train,node_cvae_train,node_vq_train,node_inference,node_loss toneRose
```

## Components

### User entry points

- [Training CLI](https://github.com/julienexr/autoencoders-lab/blob/main/main.py): `main.py`
- [Interactive demo](https://github.com/julienexr/autoencoders-lab/blob/main/app.py): `app.py`

### Data and outputs

- [Image datasets](https://github.com/julienexr/autoencoders-lab/blob/main/src/data.py): `src/data.py`
- [Model checkpoints](https://github.com/julienexr/autoencoders-lab/tree/main/models): `models`
- [Visual outputs](https://github.com/julienexr/autoencoders-lab/blob/main/src/visualization.py): `src/visualization.py`

### Autoencoder models

- [Autoencoder](https://github.com/julienexr/autoencoders-lab/blob/main/modules/autoencoder/ae.py): `modules/autoencoder/ae.py`
- [Variational AE](https://github.com/julienexr/autoencoders-lab/blob/main/modules/vae/vae.py): `modules/vae/vae.py`
- [Conditional VAE](https://github.com/julienexr/autoencoders-lab/blob/main/modules/cvae/cvae.py): `modules/cvae/cvae.py`
- [Vector-quantized AE](https://github.com/julienexr/autoencoders-lab/blob/main/modules/vq_vae/vq_vae.py): `modules/vq_vae/vq_vae.py`
- [Transformer prior](https://github.com/julienexr/autoencoders-lab/blob/main/modules/vq_vae/tranformer_prior.py): `modules/vq_vae/tranformer_prior.py`

### Training and visualization

- [AE training](https://github.com/julienexr/autoencoders-lab/blob/main/modules/autoencoder/training.py): `modules/autoencoder/training.py`
- [VAE training](https://github.com/julienexr/autoencoders-lab/blob/main/modules/vae/training.py): `modules/vae/training.py`
- [CVAE training](https://github.com/julienexr/autoencoders-lab/blob/main/modules/cvae/training.py): `modules/cvae/training.py`
- [VQ-VAE training](https://github.com/julienexr/autoencoders-lab/blob/main/modules/vq_vae/training.py): `modules/vq_vae/training.py`
- [VQ prior inference](https://github.com/julienexr/autoencoders-lab/blob/main/modules/vq_vae/inference.py): `modules/vq_vae/inference.py`
- [Perceptual loss](https://github.com/julienexr/autoencoders-lab/blob/main/src/loss.py): `src/loss.py`

### Other components

- **User**

## Connections

- User → Training CLI: chooses training (evidence: `README.md`)
- User → Interactive demo: opens demo (evidence: `README.md`)
- Training CLI → Image datasets: loads data (evidence: `main.py`)
- Training CLI → AE training: trains AE (evidence: `main.py`)
- Training CLI → VAE training: trains VAE (evidence: `main.py`)
- Training CLI → Model checkpoints: saves weights (evidence: `main.py`)
- AE training → Visual outputs: renders outputs (evidence: `modules/autoencoder/training.py`)
- VAE training → Visual outputs: renders outputs (evidence: `modules/vae/training.py`)
- VAE training → Perceptual loss: computes perceptual loss (evidence: `modules/vae/training.py`)
- CVAE training → Visual outputs: renders outputs (evidence: `modules/cvae/training.py`)
- VQ-VAE training → Visual outputs: renders outputs (evidence: `modules/vq_vae/training.py`)
- VQ-VAE training → Vector-quantized AE: encodes batches (evidence: `modules/vq_vae/training.py`)
- VQ-VAE training → Transformer prior: trains prior (evidence: `modules/vq_vae/training.py`)
- VQ prior inference → Transformer prior: samples indices (evidence: `modules/vq_vae/inference.py`)
- VQ prior inference → Vector-quantized AE: decodes samples (evidence: `modules/vq_vae/inference.py`)
- Interactive demo → Image datasets: loads examples (evidence: `app.py`)
- Interactive demo → Autoencoder: runs model (evidence: `app.py`)
- Interactive demo → Variational AE: runs model (evidence: `app.py`)
- Interactive demo → Conditional VAE: runs model (evidence: `app.py`)
- Interactive demo → Vector-quantized AE: runs model (evidence: `app.py`)
- Interactive demo → Transformer prior: uses prior (evidence: `app.py`)
- Interactive demo → Model checkpoints: loads weights
- Image datasets → User: downloads datasets (evidence: `src/data.py`)
