# akshayaa-403/imdb-sentiment-master architecture

> Architecture diagram of the GitHub repository akshayaa-403/imdb-sentiment-master, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2025-08-17). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/akshayaa-403/imdb-sentiment-master
- Repository: https://github.com/akshayaa-403/imdb-sentiment-master

## Overview

##### Project type and purpose
- This is an NLP deep-learning pipeline implemented as a Jupyter Notebook (“IMDB_Sentiment_Master.ipynb”).
- Purpose: ingest the IMDB reviews dataset, preprocess text, train a hybrid TensorFlow model (Embedding → Bi-LSTM → pooling → dense) for binary sentiment classification, evaluate performance, and potentially extend to a real-time prediction interface.

##### File structure
- Top-level files only:
- IMDB_Sentiment_Master.ipynb – the end-to-end notebook containing code for data loading, preprocessing, model definition, training, evaluation, and visualization.
- README.md – project overview, pipeline description, hyperparameters, metrics, future extensions.
- LICENSE, .gitignore.
- No separate `src/` or `scripts/` directories; logically, code is organized in sequential notebook cells.

##### README for additional insights
- Architecture section details a hybrid embedding + recurrent + convolutional model.
- Preprocessing steps are explicitly enumerated (tokenization, standardization, padding, embedding lookup, augmentation).
- Performance metrics and evaluation (accuracy, confusion matrix) are included.
- Future improvements hint at transformer integration and web deployment.

## Diagram (Mermaid)

```mermaid
flowchart TD
    subgraph "Project Documentation"
        Doc["README.md"]:::external
        License["LICENSE"]:::external
        Gitignore[".gitignore"]:::external
    end

    subgraph "Pipeline"
        DataIngest["Data Source: TF IMDB Dataset"]:::data
        Preprocessing["Preprocessing Module\n(tokenization, standardization,\npadding, embedding lookup,\naugmentation)"]:::data
        ModelDef["Model Definition\n(Embedding → SpatialDropout →\nBi-LSTM → GlobalMaxPooling →\nDense → Output)"]:::model
        Training["Training Engine\n(TF training loop, callbacks,\nearly stopping)"]:::train
        EvalVis["Evaluation & Visualization\n(accuracy, confusion matrix,\ntraining history plots)"]:::eval
        Storage["Storage\n(checkpoints & model weights)"]:::storage
    end

    Future["Serving Layer (Future)\nReal-time prediction API/UI"]:::future

    DataIngest -->|"raw text reviews"| Preprocessing
    Preprocessing -->|"token sequences\npadded arrays"| ModelDef
    ModelDef -->|"model graphs\ninput tensors"| Training
    Training -->|"metrics & logs"| EvalVis
    Training -->|"saved weights"| Storage
    Storage -->|"deployed model"| Future

    click Doc "https://github.com/akshayaa-403/imdb-sentiment-master/blob/main/README.md"
    click License "https://github.com/akshayaa-403/imdb-sentiment-master/tree/main/LICENSE"
    click Gitignore "https://github.com/akshayaa-403/imdb-sentiment-master/blob/main/.gitignore"
    click DataIngest "https://github.com/akshayaa-403/imdb-sentiment-master/blob/main/IMDB_Sentiment_Master.ipynb"
    click Preprocessing "https://github.com/akshayaa-403/imdb-sentiment-master/blob/main/IMDB_Sentiment_Master.ipynb"
    click ModelDef "https://github.com/akshayaa-403/imdb-sentiment-master/blob/main/IMDB_Sentiment_Master.ipynb"
    click Training "https://github.com/akshayaa-403/imdb-sentiment-master/blob/main/IMDB_Sentiment_Master.ipynb"
    click EvalVis "https://github.com/akshayaa-403/imdb-sentiment-master/blob/main/IMDB_Sentiment_Master.ipynb"
    click Storage "https://github.com/akshayaa-403/imdb-sentiment-master/blob/main/IMDB_Sentiment_Master.ipynb"

    classDef data fill:#ADD8E6,stroke:#000
    classDef model fill:#90EE90,stroke:#000
    classDef train fill:#FFD580,stroke:#000
    classDef eval fill:#D8BFD8,stroke:#000
    classDef storage fill:#D3D3D3,stroke:#000
    classDef future fill:#F0E68C,stroke:#000
    classDef external fill:#E0E0E0,stroke:#000
```
