| Train Model | Use Model |
|---|---|
| training a model is the process of improving how well a model works. Training requires that we use the model, as well as the objective function and optimizer, in a special loop | |
| (takes few minutes or more) | Training Model+Using trained model in real world to get output on input dataset |
| (takes less than few seconds) |
Train Model
Using Model:
To get estimations for a input dataset
| Features | columns in dataset used as input for training model |
|---|---|
feature is harness size: for avalanche-rescue dog store scenario |
|
| Label | models predicted output |
boot size is our label: for avalanche-rescue dog store scenario |
|
| Using Model | need Features(input) only |
| Training Model | need both Features(input) and Label(predicted output) |
| Save Model in File |
|---|
| We no longer need the original data, the objective function, or the model updater. |
| When we want to use the model, we can load it from disk, provide it with new data, and get back a prediction. |