Example Analysis of Keras High-level Interface Application

This article mainly introduces the example analysis of Keras high-level interface application, the article is very detailed, has a certain reference value, interested friends must read it!

1.metrics

Keras.metrics can be used to record and track data. We can use this interface when we have too much data and want to take a look at the training in the middle. The steps are as follows:

① creates Meter

Create a meter using the excuses included in metrics.

② updates data

When we get the data we need in a line of code, we can call the update_state method to update the data. Note that different methods require different parameters.

③ acquires data

We can get the data stored by the current meter when we set it to a node or state.

④ reset statu

When the viewing of one phase data record is finished, use reset_states to reset the meter to record the next phase.

two。 Quick training

After generating a model, there are compile, fit, evalute, predict and other interfaces that can be called, which makes our training easy to implement.

① compile

This method can specify: optimizer + lr, loss, accuracy, etc.

② fit

After completing the compile, call fit directly, give the training data, and specify the epoch.

The above two lines can directly complete the training, and some basic information will be returned in the training process, such as the training cycle, the amount of data, the time used, the time used in each step, the loss value after each cycle, and so on.

Fit can also give validation_data = test_data, validation_freq=2, that is, cycle 2 training, take a test, will print out the test score. It can be seen that the following situation has been over-fitted.

③ evaluate

After calling this method, the test will be carried out after the training is completed, and the test results will be printed out.

④ predict

This is actually the same as network (x), which is to complete the forward propagation.

The above is all the contents of the article "sample Analysis of Keras High-level Interface Application". Thank you for reading! Hope to share the content to help you, more related knowledge, welcome to follow the industry information channel!