Nets4Learning

Web platform for the design and execution of deep learning models for learning and initiation in the study of deep learning models.

The tool proposes different classical machine learning problems with known data sets to study and model different neural network architectures and training parameters. The tool addresses different examples of deep learning models such as tabular classification, image classifier or object identification.

There are some classical problems prepared and reviewed to make predictions, the tool has the feature to pre process data sets that the user uploads, train models and predict in which class it would be classified.

Datasets and Models

1. Tabular classification
   • MODEL_CAR.js
   • MODEL_IRIS.js
   • MODEL_LYMPHOGRAPHY.js
2. Regression
   • MODEL_1_SALARY.js
   • MODEL_2_AUTO_MPG.js
   • MODEL_3_HOUSING_PRICES.js
   • MODEL_4_BREAST_CANCER.js
   • MODEL_5_STUDENT_PERFORMANCE.js
   • MODEL_6_WINE.js
3. Image classifier
   • MODEL_IMAGE_MNIST.js
   • MODEL_IMAGE_MOBILENET.js
4. Object identification
   • MODEL_1_FACE_DETECTOR.js
   • MODEL_2_FACE_MESH.js
   • MODEL_3_MOVE_NET_POSE_NET.js
   • MODEL_4_COCO_SSD.js
   • MODEL_5_FACE_API.js
   • MODEL_6_HAND_SIGN.js

Install local

export NODE_OPTIONS="--max-old-space-size=8192"
npm install
npm start

Project environment

Create the files .env, .env.development, .env.production, .env.simidat, .env.simidat-beta, .

WATCHPACK_POLLING=true
FAST_REFRESH=true
NODE_ENV="production"
PUBLIC_URL="https://simidat.ujaen.es/n4l-beta"
REACT_APP_PUBLIC_URL="https://simidat.ujaen.es/n4l-beta"
REACT_APP_PATH="/n4l-beta"
REACT_APP_ENVIRONMENT="production"
REACT_APP_GA_MEASUREMENT_ID="G-XXXXXXXXXX"
REACT_APP_SHOW_NEW_FEATURE="true"
REACT_APP_NEW_FEATURE="linear-regression"

Deploy with Traefik and Docker

docker-compose.yml

docker compose down n4l
docker compose build n4l
docker compose up n4l

networks:
 - proxy

services:
  # traefik: ...

  n4l:
    container_name: "${PROJECT_NAME}_n4l"
    build:
      context: ./Path/To/n4l-repository/
      dockerfile: Dockerfile
      args:
        ARG_BUILD: simidat
    networks:
      - proxy
    labels:
      - "traefik.http.routers.${PROJECT_NAME}_n4l.entrypoints=https"
      - "traefik.http.routers.${PROJECT_NAME}_n4l.rule=Host(`example.com`) && PathPrefix(`/n4l`)"
      - "traefik.http.routers.${PROJECT_NAME}_n4l.tls=true"
      - "traefik.http.routers.${PROJECT_NAME}_n4l.middlewares=secure-public@file,n4l-stripprefix"
      - "traefik.http.middlewares.n4l-stripprefix.stripprefix.prefixes=/n4l"
      - "traefik.http.services.n4l.loadbalancer.server.port=3000"

Command matrix

	development	build
Development	npm start
simidat		npm run build:simidat
simidat-beta		npm run build:simidat-beta

node -v 
# v18.16.1
npm -v 
# 9.5.1