configs/data_preprocessing_c23.yaml

[userwxp]

PREPROCESSING: DATASET: FaceForensics SPLIT: train ROOT: /data/deepfake_cluster/datasets_df/FaceForensics++/c23/ FAKETYPE: [Deepfakes, Face2Face, FaceSwap, NeuralTextures] IMAGE_SUFFIX: jpg DATA_TYPE: images LABEL: [real] facial_lm_pretrained: pretrained/shape_predictor_81_face_landmarks.dat N_LANDMARKS: 81 SAMPLING: ACTIVE: False NUMBERS: 8 DEBUG: False

问题1：为什么FAKETYPE: [Deepfakes, Face2Face, FaceSwap, NeuralTextures]对应的标签是 LABEL: [real] 难道不应该是fake吗 ，
问题2：不需要original的类型的landmark吗