Model training method, attack detection method and related device

The embodiment of the invention discloses a model training method, an attack detection method and a related device, and the method comprises the steps: obtaining the contrast data of doubt data, and carrying out the training adjustment of a pre-used model through employing the labeled data, the doubt data and the contrast data, so as to obtain a target prediction model. Wherein the suspected data, the contrast data and the labeled data correspondingly marked with the affiliated attack type labels are adopted to re-train the pre-trained pre-used model to obtain the target prediction model, so that more data possibly with attack features can be found, the applicability of the model in a specific task scene is improved, and the prediction efficiency is improved. According to the method and the system, whether the to-be-detected data really has an attack behavior or not is predicted, similar black words are effectively mined, and more possible bypassing technique features are found, so that attack detection depen