Large-scale pre-training method, device and equipment for multi-modal data of power transformation main equipment

The invention discloses a large-scale pre-training method, device and equipment for multi-modal data of power transformation main equipment, and the method comprises the steps: obtaining initial multi-modal data, carrying out the preprocessing of the initial multi-modal data, obtaining the multi-modal data, and enabling the multi-modal data to comprise the online monitoring data and voiceprint data of the power transformation main equipment; respectively carrying out hidden layer feature extraction on the online monitoring quantity data and the voiceprint data, and determining a first hidden layer feature of the online monitoring quantity data and a second hidden layer feature of the voiceprint data; performing enhancement processing on the first hidden layer feature and the second hidden layer feature, and performing splicing processing to obtain a target feature; and performing multi-modal interaction processing based on the target features, constructing a loss function based on a multi-modal interaction pr