Skip to content

building models using TensorFlow’s low level API

Notifications You must be signed in to change notification settings

psun6789/Deep_Learning

Repository files navigation

Deep_Learning

Submission: Date: 04-April-2022

Assignment-01: Duilding models using TensorFlow’s low level API

Description: The Fashion-MNIST dataset is a basic image dataset consisting of a training set of 60,000 examples and a test set of 10,000 examples. Each example is a 28x28 grayscale image (784 pixel values in total), associated with a label from 10 different classes.

The following are the set of classes in this classification problem (the associated integer class label is listed in brackets). • T-shirt/top (0) • Trouser (1) • Pullover (2) • Dress (3) • Coat (4) • Sandal (5) • Shirt (6) • Sneaker (7) • Bag (8) • Ankle boot (9)

Question 1_1_1.ipynb: Network architecture A a. Layer 1: 200 neurons (ReLu activation functions). b. Layer 2: Softmax Layer

Question 1_1_2.ipynb: Network architecture B a. Layer 1: 300 neurons (ReLu activation functions). b. Layer 2: 100 neurons (ReLu activation function) c. Layer 3: Softmax Layer

Question 1_2_1.ipynb: Network architecture, DropuOut Regularisation a. Layer 1: 300 neurons (ReLu activation functions). b. Layer 2: Dropout Layer c. Layer 2: 100 neurons (ReLu activation function) d. Layer 3: Softmax Layer

Question 1_3_1.ipynb: Mini Batch a. Layer 1: 300 neurons (ReLu activation functions). b. Layer 2: Dropout Layer c. Layer 2: 100 neurons (ReLu activation function) d. Layer 3: Softmax Layer

Question 1_4_1.ipynb: Auto-Encoder, Network Architecture a. Layer 1: 128 Neurons with ReLu activations b. Layer 2: 64 Neurons with ReLu activations c. Layer 3: 32 Neurons with ReLu activations d. Layer 4: 64 Neurons with ReLu activations e. Layer 5: 128 Neurons with ReLu activations f. Layer 6: 784 Neurons with Sigmoid activations

About

building models using TensorFlow’s low level API

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published