# rahulrajln1111/dgcnn architecture

> Architecture diagram of the GitHub repository rahulrajln1111/dgcnn, 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/rahulrajln1111/dgcnn
- Repository: https://github.com/rahulrajln1111/dgcnn

## Overview

This repository combines an edge-deployable DGCNN classifier with a separate, experimental hardware-aware architecture-search workflow. The main user path is to obtain ModelNet10 data, train a configurable classifier, then evaluate or benchmark a checkpoint on edge hardware. Additional capabilities include model profiling, ablation sweeps, and result visualization. The `nouse/` tree contains the HGNAS search system. Some of its source was not sampled; its documented module relationships are included only where the SOURCE INDEX or excerpts support them.

Main entry points: [`nouse/main.py`](https://github.com/rahulrajln1111/dgcnn/blob/master/nouse/main.py) (Search Orchestrator), [`nouse/search.py`](https://github.com/rahulrajln1111/dgcnn/blob/master/nouse/search.py) (Evolutionary Search), [`dataset.py`](https://github.com/rahulrajln1111/dgcnn/blob/master/dataset.py) (Point-Cloud Dataset), [`train_ablation.py`](https://github.com/rahulrajln1111/dgcnn/blob/master/train_ablation.py) (Ablation Sweeps)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_classifier["DGCNN Classifier"]
  node_dataset["Point-Cloud Dataset<br/>[dataset.py]"]
  node_model["DGCNN Model<br/>[dgcnn_model.py]"]
  node_train["Training Loop<br/>[train.py]"]
  node_download["Data Fetcher<br/>[download_data.py]"]
end

subgraph group_deploy["Edge Evaluation"]
  node_inference["Test Inference<br/>[inference.py]"]
  node_benchmark["Hardware Benchmark<br/>[benchmark.py]"]
  node_profile["Layer Profiling<br/>[profile_model.py]"]
end

subgraph group_experiments["Analysis"]
  node_ablation["Ablation Sweeps<br/>[train_ablation.py]"]
  node_experiment_cfg["Experiment Configurations<br/>[experiments.py]"]
  node_visualize["Result Visualization<br/>[visualize.py]"]
end

subgraph group_nas["HGNAS Search"]
  node_nas_main["Search Orchestrator<br/>[main.py]"]
  node_nas_config["Search Configuration<br/>[config.py]"]
  node_nas_data["NAS Data Loading<br/>[data_loader.py]"]
  node_design["Architecture Space<br/>[design_space.py]"]
  node_supernet["GNN Supernet<br/>[supernet.py]"]
  node_trainer["Supernet Training<br/>[trainer.py]"]
  node_search["Evolutionary Search<br/>[search.py]"]
  node_predictor["Hardware Predictor<br/>[hw_predictor.py]"]
  node_memory["Memory Estimator"]
  node_plot_res["Search Result Plots<br/>[plot_res.py]"]
end

node_user(("User"))
node_modelnet[("ModelNet10 Files")]

node_user -->|"fetches data"| node_download
node_download -->|"downloads"| node_modelnet
node_modelnet -->|"provides files"| node_dataset
node_user -->|"trains model"| node_train
node_train -->|"loads samples"| node_dataset
node_train -->|"optimizes"| node_model
node_user -->|"evaluates checkpoint"| node_inference
node_inference -->|"loads test data"| node_dataset
node_inference -->|"runs inference"| node_model
node_user -->|"benchmarks device"| node_benchmark
node_user -->|"profiles layers"| node_profile
node_profile -->|"profiles"| node_model
node_profile -->|"loads data"| node_dataset
node_user -->|"runs sweeps"| node_ablation
node_ablation -->|"trains variants"| node_model
node_ablation -->|"loads samples"| node_dataset
node_ablation -->|"uses settings"| node_experiment_cfg
node_user -->|"visualizes results"| node_visualize
node_nas_main -->|"loads data"| node_nas_data
node_nas_main -->|"initializes"| node_supernet
node_nas_main -->|"sets search space"| node_design
node_nas_main -->|"trains predictor"| node_predictor
node_nas_main -->|"trains supernet"| node_trainer
node_nas_main -->|"runs search"| node_search
node_nas_main -->|"plots results"| node_plot_res
node_search -->|"mutates architectures"| node_design
node_search -->|"estimates hardware cost"| node_predictor
node_search -->|"evaluates candidates"| node_trainer
node_search -->|"scores candidates"| node_supernet
node_trainer -->|"trains paths"| node_supernet
node_trainer -->|"samples architectures"| node_design
node_predictor -->|"encodes architectures"| node_design
node_predictor -->|"uses estimates"| node_memory

click node_dataset "https://github.com/rahulrajln1111/dgcnn/blob/master/dataset.py"
click node_model "https://github.com/rahulrajln1111/dgcnn/blob/master/dgcnn_model.py"
click node_train "https://github.com/rahulrajln1111/dgcnn/blob/master/train.py"
click node_download "https://github.com/rahulrajln1111/dgcnn/blob/master/download_data.py"
click node_inference "https://github.com/rahulrajln1111/dgcnn/blob/master/inference.py"
click node_benchmark "https://github.com/rahulrajln1111/dgcnn/blob/master/benchmark.py"
click node_profile "https://github.com/rahulrajln1111/dgcnn/blob/master/profile_model.py"
click node_ablation "https://github.com/rahulrajln1111/dgcnn/blob/master/train_ablation.py"
click node_experiment_cfg "https://github.com/rahulrajln1111/dgcnn/blob/master/experiments.py"
click node_visualize "https://github.com/rahulrajln1111/dgcnn/blob/master/visualize.py"
click node_nas_main "https://github.com/rahulrajln1111/dgcnn/blob/master/nouse/main.py"
click node_nas_config "https://github.com/rahulrajln1111/dgcnn/blob/master/nouse/config.py"
click node_nas_data "https://github.com/rahulrajln1111/dgcnn/blob/master/nouse/data_loader.py"
click node_design "https://github.com/rahulrajln1111/dgcnn/blob/master/nouse/design_space.py"
click node_supernet "https://github.com/rahulrajln1111/dgcnn/blob/master/nouse/supernet.py"
click node_trainer "https://github.com/rahulrajln1111/dgcnn/blob/master/nouse/trainer.py"
click node_search "https://github.com/rahulrajln1111/dgcnn/blob/master/nouse/search.py"
click node_predictor "https://github.com/rahulrajln1111/dgcnn/blob/master/nouse/hw_predictor.py"
click node_memory "https://github.com/rahulrajln1111/dgcnn/blob/master/nouse/memory_estimator.py"
click node_plot_res "https://github.com/rahulrajln1111/dgcnn/blob/master/nouse/plot_res.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_dataset,node_model,node_train,node_download,node_user toneBlue
class node_inference,node_benchmark,node_profile,node_modelnet toneAmber
class node_ablation,node_experiment_cfg,node_visualize toneMint
class node_nas_main,node_nas_config,node_nas_data,node_design,node_supernet,node_trainer,node_search,node_predictor,node_memory,node_plot_res toneRose
```

## Components

### DGCNN Classifier

- [Point-Cloud Dataset](https://github.com/rahulrajln1111/dgcnn/blob/master/dataset.py): `dataset.py`
- [DGCNN Model](https://github.com/rahulrajln1111/dgcnn/blob/master/dgcnn_model.py): `dgcnn_model.py`
- [Training Loop](https://github.com/rahulrajln1111/dgcnn/blob/master/train.py): `train.py`
- [Data Fetcher](https://github.com/rahulrajln1111/dgcnn/blob/master/download_data.py): `download_data.py`

### Edge Evaluation

- [Test Inference](https://github.com/rahulrajln1111/dgcnn/blob/master/inference.py): `inference.py`
- [Hardware Benchmark](https://github.com/rahulrajln1111/dgcnn/blob/master/benchmark.py): `benchmark.py`
- [Layer Profiling](https://github.com/rahulrajln1111/dgcnn/blob/master/profile_model.py): `profile_model.py`

### Analysis

- [Ablation Sweeps](https://github.com/rahulrajln1111/dgcnn/blob/master/train_ablation.py): `train_ablation.py`
- [Experiment Configurations](https://github.com/rahulrajln1111/dgcnn/blob/master/experiments.py): `experiments.py`
- [Result Visualization](https://github.com/rahulrajln1111/dgcnn/blob/master/visualize.py): `visualize.py`

### HGNAS Search

- [Search Orchestrator](https://github.com/rahulrajln1111/dgcnn/blob/master/nouse/main.py): `nouse/main.py`
- [Search Configuration](https://github.com/rahulrajln1111/dgcnn/blob/master/nouse/config.py): `nouse/config.py`
- [NAS Data Loading](https://github.com/rahulrajln1111/dgcnn/blob/master/nouse/data_loader.py): `nouse/data_loader.py`
- [Architecture Space](https://github.com/rahulrajln1111/dgcnn/blob/master/nouse/design_space.py): `nouse/design_space.py`
- [GNN Supernet](https://github.com/rahulrajln1111/dgcnn/blob/master/nouse/supernet.py): `nouse/supernet.py`
- [Supernet Training](https://github.com/rahulrajln1111/dgcnn/blob/master/nouse/trainer.py): `nouse/trainer.py`
- [Evolutionary Search](https://github.com/rahulrajln1111/dgcnn/blob/master/nouse/search.py): `nouse/search.py`
- [Hardware Predictor](https://github.com/rahulrajln1111/dgcnn/blob/master/nouse/hw_predictor.py): `nouse/hw_predictor.py`
- [Memory Estimator](https://github.com/rahulrajln1111/dgcnn/blob/master/nouse/memory_estimator.py): `nouse/memory_estimator.py`
- [Search Result Plots](https://github.com/rahulrajln1111/dgcnn/blob/master/nouse/plot_res.py): `nouse/plot_res.py`

### Other components

- **User**
- **ModelNet10 Files**

## Connections

- User → Data Fetcher: fetches data (evidence: `README.md`)
- Data Fetcher → ModelNet10 Files: downloads (evidence: `README.md`)
- ModelNet10 Files → Point-Cloud Dataset: provides files (evidence: `dataset.py`)
- User → Training Loop: trains model (evidence: `README.md`)
- Training Loop → Point-Cloud Dataset: loads samples (evidence: `train.py`)
- Training Loop → DGCNN Model: optimizes (evidence: `train.py`)
- User → Test Inference: evaluates checkpoint (evidence: `README.md`)
- Test Inference → Point-Cloud Dataset: loads test data (evidence: `inference.py`)
- Test Inference → DGCNN Model: runs inference (evidence: `inference.py`)
- User → Hardware Benchmark: benchmarks device (evidence: `README.md`)
- User → Layer Profiling: profiles layers (evidence: `README.md`)
- Layer Profiling → DGCNN Model: profiles (evidence: `profile_model.py`)
- Layer Profiling → Point-Cloud Dataset: loads data (evidence: `profile_model.py`)
- User → Ablation Sweeps: runs sweeps (evidence: `README.md`)
- Ablation Sweeps → DGCNN Model: trains variants (evidence: `train_ablation.py`)
- Ablation Sweeps → Point-Cloud Dataset: loads samples (evidence: `train_ablation.py`)
- Ablation Sweeps → Experiment Configurations: uses settings (evidence: `train_ablation.py`)
- User → Result Visualization: visualizes results (evidence: `README.md`)
- Search Orchestrator → NAS Data Loading: loads data (evidence: `nouse/main.py`)
- Search Orchestrator → GNN Supernet: initializes (evidence: `nouse/main.py`)
- Search Orchestrator → Architecture Space: sets search space (evidence: `nouse/main.py`)
- Search Orchestrator → Hardware Predictor: trains predictor (evidence: `nouse/main.py`)
- Search Orchestrator → Supernet Training: trains supernet (evidence: `nouse/main.py`)
- Search Orchestrator → Evolutionary Search: runs search (evidence: `nouse/main.py`)
- Search Orchestrator → Search Result Plots: plots results (evidence: `nouse/main.py`)
- Evolutionary Search → Architecture Space: mutates architectures (evidence: `nouse/search.py`)
- Evolutionary Search → Hardware Predictor: estimates hardware cost (evidence: `nouse/search.py`)
- Evolutionary Search → Supernet Training: evaluates candidates (evidence: `nouse/search.py`)
- Evolutionary Search → GNN Supernet: scores candidates (evidence: `nouse/search.py`)
- Supernet Training → GNN Supernet: trains paths (evidence: `nouse/trainer.py`)
- Supernet Training → Architecture Space: samples architectures (evidence: `nouse/trainer.py`)
- Hardware Predictor → Architecture Space: encodes architectures (evidence: `nouse/hw_predictor.py`)
- Hardware Predictor → Memory Estimator: uses estimates (evidence: `nouse/hw_predictor.py`)
