1) Implement the Convolution neural networks (CNN) approach on Fashion M-Nist dataset by importing the data from Keras library. The main objective of the assignment is to increase the accuracy of your entire model >90%.
a. Execute the CNN approach to train your model to classify different labels that are present in the Fashion-Mnist dataset. Implement Normalization technique to achieve zero mean and zero variance and One-hot encoding if required. Comment on your results.
b. Design Convolution neural networks with necessary layers to achieve the accurate model for classifying the different classes.
i. Report all the layers that you considered in your model and illustrate the functionality of each layer. Report the results obtained for various learning rates that you considered. Also, illustrate all the hyper parameters (e.g., number of filters, Dimensions of Filter, Stride, Padding, Activation Function, etc.). Make sure to elaborate the reason for considering specific parameters with mathematical calculation wherever required.
ii. Show the type of optimizers used out of various options during the compilation of the model and type of cross entropy for classifying multi classes. Decay the learn rate. Describe your parameter decay approach and comment on your results training on number of epochs.
iii. Plot the graph for training and validation curves. Report the classification report (e.g., F1_score, precision, Recall, contingency table etc.).
iv. Provide a report to validate that your model has good predicting capability over the classes
deadline - 20th April
13 freelanceria on tarjonnut keskimäärin $57 tähän työhön
Hey! I am having 4+ years of Industry Experience in Machine Learning, Deep Learning,Natural Language Processing, and Computer Vision Applications. Message me to discuss more details
Hello, I'm python developper and data scientist with 3+ years of experience. I've been through your assignment and I'm sure I can deliver you good work with good explanation.