Model-release-v0.0.1

Improvements:

• More data for training
• Remove 'lr' = 0.001 as it causes problems
• Changed optimiser from RMSProp to Adam
• Changed from "<=" to ">" threshold for determining whether the prediction is true or false