Modular reinforcement learning model processing method, system and equipment and storage medium

Abstract

The invention relates to a modular reinforcement learning model processing method and device, computer equipment and a storage medium. The method comprises that: interaction data generated by a virtual object in an interaction process with an interaction environment is obtained; the virtual object is controlled by a running component in a reinforcement learning system deployed in the cloud; the reinforcement learning system further comprises a learning assembly and an evaluation assembly; the reinforcement learning model is iteratively trained based on the interaction data through a learning component; in the iterative training process, the reinforcement learning model obtained through iterative training is evaluated through an evaluation component, and whether the reinforcement learning model obtained through iterative training meets interaction conditions or not is judged according to a result obtained through evaluation; and if not, the model associated with the running component is updated according to the reinforcement learning model obtained by iterative training, so that the running component controls the virtual object based on the updated reinforcement learning model. By adopting the method, the complexity of multiplexing the reinforcement learning model training framework in different services can be reduced.

Description

technical field [0001] The present application relates to the technical field of artificial intelligence, in particular to a model training method, device, computer equipment and storage medium. Background technique [0002] With the development of artificial intelligence technology, reinforcement learning technology is widely used in games, e-commerce recommendation, automatic driving, intelligent scheduling and other fields. The training process of the reinforcement learning model is complex, how to train the reinforcement learning model is an important issue to promote the development of reinforcement learning technology. In the traditional technology, the open source framework of RLLib (Reinforcement Learning Library) distributed reinforcement learning is used to train the reinforcement learning model. However, the various parts of the RLLib framework are interrelated. When the reinforcement learning model training process based on the RLLib framework is reused in vario...

Images