# somepago/saint architecture

> Architecture diagram of the GitHub repository somepago/saint, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2024-12-30). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/somepago/saint
- Repository: https://github.com/somepago/saint

## Overview

##### Main Components

a) Data Processing Layer
- Input handling for OpenML datasets
- Augmentation pipeline (from augmentations.py)
- Data preprocessing components

b) Model Architecture Layer
- Core SAINT model components:
- Column Attention (SAINT-s)
- Row Attention (SAINT-i)
- Combined Attention (SAINT)
- Embedding components for:
- Continuous data
- Categorical data
- Transformer layers with configurable depth

c) Training Pipeline
- Main training flow
- Pretraining components
- Contrastive learning module
- Loss computation units

##### Component Relationships

**Data Input → Preprocessing → Model → Training/Evaluation → Output**

b) Parallel paths for:
- Regular training pipeline
- Pretraining pipeline (when enabled)

##### Specific Elements

a) Input Layer:
- OpenML dataset interface
- Data augmentation modules (mixup, CutMix)
- Data type handling (continuous/categorical)

b) Model Core:
- Embedding layer
- Transformer blocks
- Projection heads for contrastive learning

c) Training Components:
- Loss computation units
- Evaluation metrics (AuROC, Accuracy, RMSE)
- Model selection based on validation performance

##### Important Details to Highlight

- Configurable components (embedding size, transformer depth, attention heads)
- Multiple task support (regression, binary, multiclass)
- Pretraining pathway activation
- Semi-supervised learning capabilities

##### Technical Specifications to Note

- Framework: PyTorch
- GPU Support: NVIDIA RTX 2080Ti
- Python version: 3.8
- Key dependencies (torch, torchvision)

## Diagram (Mermaid)

```mermaid
graph TB
    subgraph Data_Processing
        DI[Data Input]
        OML["OpenML Dataset Handler"]:::data
        AUG["Augmentation Pipeline"]:::data
        DP["Data Preprocessing"]
        DI --> OML
        OML --> AUG
        AUG --> DP
    end

    subgraph Model_Architecture
        subgraph Embedding
            CE["Continuous Embedding"]
            CAE["Categorical Embedding"]
        end

        subgraph Attention_Mechanisms
            CA["Column Attention (SAINT-s)"]
            RA["Row Attention (SAINT-i)"]
            CRA["Combined Attention (SAINT)"]
        end

        subgraph Model_Variants
            CM["Core SAINT Model"]:::model
            PM["Pretraining Model"]:::model
            VPM["Vision Pretraining Model"]:::model
        end
    end

    subgraph Training_Pipeline
        MT["Main Training"]:::train
        RT["Robust Training"]:::train
        PT["Pretraining"]:::train
        UTIL["Utility Functions"]:::util
        
        subgraph Training_Components
            LC["Loss Computation"]
            EM["Evaluation Metrics"]
            MS["Model Selection"]
        end
    end

    subgraph Legacy_System
        OM["Old Model"]:::legacy
        OD["Old Data Processing"]:::legacy
        OT["Old Training"]:::legacy
        OP["Old Pretraining"]:::legacy
    end

    %% Connections
    DP --> CE
    DP --> CAE
    CE --> CM
    CAE --> CM
    CM --> MT
    CM --> RT
    CM --> PT
    MT --> MS
    RT --> MS
    PT --> MS

    %% Styles
    classDef data fill:#90EE90
    classDef model fill:#ADD8E6
    classDef train fill:#FFA500
    classDef util fill:#DDA0DD
    classDef legacy fill:#D3D3D3

    %% Click Events
    click OML "https://github.com/somepago/saint/blob/main/data_openml.py"
    click AUG "https://github.com/somepago/saint/blob/main/augmentations.py"
    click CM "https://github.com/somepago/saint/blob/main/models/model.py"
    click PM "https://github.com/somepago/saint/blob/main/models/pretrainmodel.py"
    click VPM "https://github.com/somepago/saint/blob/main/models/pretrainmodel_vision.py"
    click MT "https://github.com/somepago/saint/blob/main/train.py"
    click RT "https://github.com/somepago/saint/blob/main/train_robust.py"
    click PT "https://github.com/somepago/saint/blob/main/pretraining.py"
    click UTIL "https://github.com/somepago/saint/blob/main/utils.py"
    click OM "https://github.com/somepago/saint/blob/main/old_version/models/model.py"
    click OD "https://github.com/somepago/saint/blob/main/old_version/data.py"
    click OT "https://github.com/somepago/saint/blob/main/old_version/train.py"
    click OP "https://github.com/somepago/saint/blob/main/old_version/train_pt.sh"
```
