How ayushbadwaik/neual-network-deep-learning-tae works
EarlyStop-Bench is a neural-network benchmark with a Flask dashboard and a standalone CLI. Users choose a benchmark or upload CSV data, configure training, and compare each architecture trained with and without early stopping; results include accuracy, loss histories, overfitting gaps, epochs, and time. The shared engine prepares datasets, constructs models, and orchestrates the comparison. The README also documents CLI charts and a CSV summary. The web excerpt confirms training and upload routes; report-export details are not sampled, so that capability is not wired separately here.
Main entry points: app.py Flask API · trainer.py Comparison Orchestrator · main.py CLI Runner · app.js Dashboard Logic
Components 15
Web Dashboard
- Dashboard UI
templates/index.html - Dashboard Logic
static/js/app.js - Flask API
app.py - Web Metrics And Charts
static/js/app.js
Dataset Preparation
- Built-in Datasets
ml_engine/dataset_factory.py - CSV Parsing
ml_engine/dataset_factory.py - Split And Scale
ml_engine/dataset_factory.py
Benchmark Engine
- Comparison Orchestrator
ml_engine/trainer.py - Model Factory
ml_engine/models.py - Neural Architectures
ml_engine/models.py - Early Stopping
ml_engine/early_stopping.py
CLI Benchmark
- CLI Runner
main.py - CLI Tables And Plots
main.py - Benchmark Outputs
Other components
- User
Connections 17
- User → Dashboard UI: opens
- Dashboard UI → Dashboard Logic: loads
- Dashboard Logic → Flask API: requests training
- Flask API → Built-in Datasets: loads dataset
- Dashboard Logic → Flask API: uploads CSV
- Flask API → CSV Parsing: parses upload
- Built-in Datasets → Split And Scale: prepares data
- CSV Parsing → Split And Scale: prepares data
- Flask API → Comparison Orchestrator: runs comparison
- CLI Runner → Built-in Datasets: loads dataset
- CLI Runner → Comparison Orchestrator: runs benchmark
- Comparison Orchestrator → Model Factory: creates model
- Model Factory → Neural Architectures: instantiates
- Comparison Orchestrator → Early Stopping: checks validation
- Flask API → Web Metrics And Charts: returns results
- CLI Runner → CLI Tables And Plots: formats results
- CLI Tables And Plots → Benchmark Outputs: writes summary and charts