With the development of information technology, teaching reform has also undergone major changes. e traditional college physical education teaching method cannot meet the needs of the majority of students, and the physical education teaching mode continues to be reformed. Microcourse is the most intuitive form of deep integration of information technology and physical education. From the perspective of the ipped classroom (FC), the physical education model has gradually changed from teacher centered to student centered. Deep learning (DL) emphasizes that learners have the ability to actively construct knowledge, e ectively transfer knowledge, and solve real problems. is design applies DL and convolutional neural network to the teaching design of physical gymnastics in colleges and universities.e application of the DL teaching model based on FC in the microcourse teaching of gymnastics in colleges and universities is studied and evaluated. e results show that the current utilization of microcourse teaching resources is too low. Interest-oriented teaching microcourses cannot improve students' interests. e proportion of students who are interested is relatively small, and more than 50% of students are not interested. Teachers generally believe that the current gymnastics microcourse needs further optimization and improvement. e poor quality of microvideos and the lack of supervision and reward mechanism in the course are the main reasons for the insu cient students' interest. e complexity of the videos and the liveliness of the discussions are the main problems of low resource utilization. e student's interest in learning is greatly improved after the application of the designed model, and the proportion increases to 82.4%. e e ect on ordinary college students is the most obvious, and the e ect of microvideo learning has been signi cantly promoted. Design mode has the most obvious improvement in improving learning e ciency and autonomous learning ability. e improvement of learning ability has increased from 18% to 72%, and the improvement of learning e ciency has increased from 39% to 82%. Meanwhile, students' interest in learning is stimulated, and the utilization of resources is improved.

This paper examines the problem of athletes’ training in sports, exploring the methods and means by which athletes can perform difficult movements in which they normally make minor training errors in order to achieve better competition results and placements. To this end, we test the explanatory and predictive effects of a theoretical model starting with planned behaviour and then use exercise planning, self-efficacy, and support as variables to develop a partial least squares regression model of sports to improve the explanation and prediction of sporting athletes’ intentions and behaviour. An improved RBF network-based method for player behaviour prediction is proposed. On the basis of the RBF analysis, the number of layers and the number of neurons in the hidden layer of the network are adjusted and optimised, respectively, to improve its generalisation and learning abilities, and the athlete behaviour prediction model is given. The results demonstrate the advantages of the improved algorithm, which in turn provides a more scientific approach to the current basketball training.