My Blog

Author: saqibkhan

  • Use Model.summary() and plot_model for Insights

    • Check the architecture and layer-wise summary of your model with model.summary().pythonCopy codefrom tensorflow.keras.utils import plot_model plot_model(model, to_file='model.png', show_shapes=True, show_layer_names=True)

  • Save and Load Models

    • You can save your model during and after training:pythonCopy codemodel.save('model.h5') # Loading the model from tensorflow.keras.models import load_model model = load_model('model.h5')

  • Custom Loss Functions and Metrics

    • You can define your own loss function or metrics to suit specific tasks:pythonCopy codeimport tensorflow.keras.backend as K def custom_loss(y_true, y_pred): return K.mean(K.square(y_pred - y_true)) model.compile(optimizer='adam', loss=custom_loss, metrics=['accuracy'])

  • MongoDB Select Record

    The findOne() method is used to select a single data from a collection in MongoDB. This method returns the first record of the collection.

    Example

    (Select Single Record)

    Select the first record from the ?employees? collection.

    Create a js file named “select.js”, having the following code:

    var http = require('http');  

var MongoClient = require('mongodb').MongoClient;  

var url = "mongodb://localhost:27017/MongoDatabase";  

MongoClient.connect(url, function(err, db) {  

  if (err) throw err;  

  db.collection("employees").findOne({}, function(err, result) {  

    if (err) throw err;  

    console.log(result.name);  

    db.close();  

  });  

});  

      Open the command terminal and run the following command:

      Node select.js  
      Node.js Select record 1

      Select Multiple Records

      The find() method is used to select all the records from collection in MongoDB.

      Example

      Select all the records from “employees” collection.

      Create a js file named “selectall.js”, having the following code:

      var MongoClient = require('mongodb').MongoClient;  

var url = "mongodb://localhost:27017/MongoDatabase";  

MongoClient.connect(url, function(err, db) {  

  if (err) throw err;  

  db.collection("employees").find({}).toArray(function(err, result) {  

    if (err) throw err;  

    console.log(result);  

    db.close();  

  });  

});  

        Open the command terminal and run the following command:

        Node selectall.js  
        Node.js Select record 2

      1. Transfer Learning

        • If you don’t have enough data, consider using transfer learning with a pretrained model.
          • Freeze the layers of the base model:pythonCopy codefor layer in base_model.layers: layer.trainable = False

      2. MongoDB Insert Record

        The insertOne method is used to insert record in MongoDB’s collection. The first argument of the insertOne method is an object which contains the name and value of each field in the record you want to insert.

        Example

        (Insert Single record)

        Insert a record in “employees” collection.

        Create a js file named “insert.js”, having the following code:

        var MongoClient = require('mongodb').MongoClient;  

var url = "mongodb://localhost:27017/ MongoDatabase";  

MongoClient.connect(url, function(err, db) {  

if (err) throw err;  

var myobj = { name: "Ajeet Kumar", age: "28", address: "Delhi" };  

db.collection("employees").insertOne(myobj, function(err, res) {  

if (err) throw err;  

console.log("1 record inserted");  

db.close();  

});  

}); 

          Open the command terminal and run the following command:

          Node insert.js  
          Node.js Insert record 1

          Now a record is inserted in the collection.

          Insert Multiple Records

          You can insert multiple records in a collection by using insert() method. The insert() method uses array of objects which contain the data you want to insert.

          Example

          Insert multiple records in the collection named “employees”.

          Create a js file name insertall.js, having the following code:

          var MongoClient = require('mongodb').MongoClient;  

var url = "mongodb://localhost:27017/ MongoDatabase";  

MongoClient.connect(url, function(err, db) {  

if (err) throw err;  

var myobj = [     

{ name: "Mahesh Sharma", age: "25", address: "Ghaziabad"},  

{ name: "Tom Moody", age: "31", address: "CA"},  

{ name: "Zahira Wasim", age: "19", address: "Islamabad"},  

{ name: "Juck Ross", age: "45", address: "London"}  

];  

db.collection("customers").insert(myobj, function(err, res) {  

if (err) throw err;  

console.log("Number of records inserted: " + res.insertedCount);  

db.close();  

});  

});  

            Open the command terminal and run the following command:

            Node insertall.js  
            Node.js Insert record 2

          1. Experiment with Optimizers

            • The choice of optimizer can impact your model’s performance. Besides Adam, try using RMSprop, SGD, or Nadam.pythonCopy codefrom tensorflow.keras.optimizers import RMSprop model.compile(optimizer=RMSprop(learning_rate=0.001), loss='categorical_crossentropy',

          2. Use TensorBoard for Visualization

            • TensorBoard helps you visualize the training process, track loss, and monitor metrics.pythonCopy codefrom tensorflow.keras.callbacks import TensorBoard tensorboard = TensorBoard(log_dir='./logs') model.fit(X_train, y_train, validation_data=(X_val, y_val), epochs=50, callbacks=[tenso

          3. MongoDB Create Collection

            MongoDB is a NoSQL database so data is stored in collection instead of table. createCollection method is used to create a collection in MongoDB.

            Example

            Create a collection named “employees”.

            Create a js file named “employees.js”, having the following data:

            var MongoClient = require('mongodb').MongoClient;  

var url = "mongodb://localhost:27017/ MongoDatabase";  

MongoClient.connect(url, function(err, db) {  

if (err) throw err;  

db.createCollection("employees", function(err, res) {  

if (err) throw err;  

console.log("Collection is created!");  

db.close();  

});  

}); 

              Open the command terminal and run the following command:

              Node employees.js  
              Node.js Create collection 1

            1. Optimize Batch Size and Learning Rate

              • Batch size and learning rate significantly affect model performance. Tune these parameters carefully.
              • Start with a small learning rate and experiment with different batch sizes (commonly 32, 64, or 128).