Skip to content

Latest commit

 

History

History
23 lines (19 loc) · 493 Bytes

File metadata and controls

23 lines (19 loc) · 493 Bytes

Things we could try

  • Changing Model Set up
    • Kernel size
    • Filters
    • Strides
    • Number of convolutional vs dense
    • Adding drop outs
    • Adding batch normization
    • Adding leaky relu
    • Loss function
    • Activtion
  • Changing input
    • Given as 254 x 254 x 3
    • Could make grey scale 254 x 254 x 1
    • Could try and downsample image 127 x 127 x 3
  • Fit parameters
    • Epochs
    • Batch size
    • Train Validation split currenly 0.2 on the training data could try other splits