user@devops:~$ cat README.md
Keras Sequential — Your First Neural Network
# Description
Your first neural network with Keras Sequential API and TensorFlow on the MNIST dataset (70,000 handwritten digits). Build a 3-layer Dense network with ReLU and Softmax activations, compile with Adam optimizer and sparse_categorical_crossentropy loss, train for 10 epochs with validation, and achieve ~97-98% test accuracy. Learn Deep Learning fundamentals: pixel normalization, 2D image flattening to vectors, epochs and batches concepts, learning curves, and confusion matrix for multiclass classification.
# Key features
$ Sequential neural network with 3 Dense layers: Input(784) → Dense(128, ReLU) → Dense(64, ReLU) → Dense(10, Softmax)
$ MNIST dataset: 70,000 handwritten digits (0-9)
$ Preprocessing: [0,1] normalization and 28×28 → 784 flattening
$ Compilation with Adam optimizer and sparse_categorical_crossentropy loss
$ Training with epochs, batches, and validation data
$ Evaluation: ~97-98% test accuracy, confusion matrix
$ Learning curves for overfitting detection
# Gallery
# Technologies used
