ICE 9 - AvniM/Big-Data-Analytics GitHub Wiki

Link to source code

Team Members:

Name: Avni Mehta, Class Id: 11

Name: (Trinadha) Raji Muppala, Class Id: 14

Task 1:

Run the CNN model with AdamOptimizer and AdagradOptimizer with filters 32,64,1024 and 16,36,128

Report the accuracy. Show the tensorboard

Input:

Project dataset: MS COCO data , interiors - bedroom, bathroom, kitchen, living . Converted the data to ubyte format

https://github.com/AvniM/Big-Data-Analytics/tree/master/Lab/Lab4/Source/JPG-PNG-to-MNIST-NN-Format-master

Code:

Have used the same code provided in the use case.

Modified the filters to 16,36,128 - 89% for the project data

Output:

Accuray with 32,64,1024 filters AdamOptimizer

Accuray with 16,36,128 filters AdamOptimizer

Accuray with 32,64,1024 filters AdagradOptimizer

Accuray with 16,36,128 filters AdagradOptimizer

Cross-Entropy:

Observation:

CNN with 32,64,1024 with 500 Epoch got good accuracy for project data

Contritbution of Team Members:

100% for both since both team members completed the assignment individually.