Model & data for automated classification of male and female house wren song

These files include deep learning algorithms and the data that are part of a publication aimed at locating and classifying male and female house wren songs in audio recordings. The data include:
WREN-SPOT_BINARY_3rdTraining_v1_Len625.pk = The final binary model for locating and classifying male and female house wren songs in audio recordings.WREN-SPOT_MULTICLASS_3rdTraining_v1_Len625.pk = The final multiclass model for locating and classifying male and female house wren songs in audio recordingsselection-tables_predicted_Final-3rd-model-evaluation = Selection tables created from the model's predictions of male and female house wren song. These are to compare to the manual selection tables for evaluating the 3rd training of the house wren male-female song binary and multiclass models. They can be opened in Raven Bioacoustics Software or the warbleR R Package.Selection tables - manual evaluation & verification = Selection tables created by manually scanning and verifying locations of male and female house wrens in audio recordings by three independent observers. These files can be opened in Raven Bioacoustics Software or the warbleR R Packagelog files for evaluation of trained models.zip = Log files of final multiclass model prediction. This is the model output from the final round of prediction for the house wren male & female song model. This was then used to create the model prediction & evaluation selection tables.All final training data: audio clips of male house wren song, female house wren song, and background noise used for model training.A sunrise time schedule (evaluation_soundfile_sunrise_times.xlsx) for adjusting file offset for data collection to the three hours following sunrise.