vgg16

1. Dataset

   1. Open sourced
   2. In house collected data

2. Model -> vgg16 (classification)

3. Training

   1. feeding data
   2. network
   3. Loss function
   4. accuracy metrics

4. Network

   1. input -> 784
   2. layer 1 -> 784
   3. layer 2 -> 784
   4. layer 3 -> 784
   5. logit -> 10

feed data from a dataset -> train model -> inference / predict