Conv2D Layer Args: filters: int, the output channels of the conv2d layer kernel_size: list, the kernel size of the conv2d layer

set logging file and format Args: filename: str, full path of the logger file to write logger_name: str, the logger name, e.g., 'm

Train the model Args: epochs: int, the number of epochs to train the model ckpt_epoch: int, the number of epochs to train the mode

Loss Function Args: y: tensor, the ground truth output: tensor,[pred , emotion_input] the output of the model