| serial_learner | |
| header Serial Pipeline | |
| title Serial Learner | |
| |#99CCCC|serial_controller| | |
| |#CCCCFF|policy| | |
| |#99CCFF|learner| | |
| |#99CCCC|serial_controller| | |
| start | |
| :init learner, \nset its learn_mode policy; | |
| |#99CCFF|learner| | |
| :get data from buffer; | |
| |#CCCCFF|policy| | |
| :data forward; | |
| :loss backward; | |
| :optimizer step, gradient update; | |
| |#99CCFF|learner| | |
| :update train info(loss, value) and log; | |
| :update learn info(iteration, priority); | |
| stop | |