Detecting anomalies in system logs with a compact convolutional transformer - Data

<p><strong>Detecting anomalies in system logs with a compact convolutional transformer - Data</strong></p> <p>Preprocessed data and a pre-trained model for the Larisch, Vitay, Hamker (2022) publication.</p> <p>The <em>data</em> directory contains the Blue Gene/L data set, after tokenization, shuffling, and splitting in a training and test set (BGL_masked_Xtrain.npy and BGL_masked_Xtest.npy, respectively) and the corresponding labels.<br> Additionally, the BGL_masked_Xtest_uniq.npy and BGL_masked_Ytest_uniq.npy contains the test data, where samples from the training set are removed.</p> <p>The <em>model</em> directory contains a pre-trained compact convolutional transformer (CCT) model. The CCT is trained on the proposed BGL training data and uses a 4x4 convolutional kernel.</p>