the first 50000 samples of MNIST will be used for training, the next 10000 for validation, and the last 10000 for testing. 1. Logistic regression with L2 penalty term with C= 10%, n=-10:1:10. For each C, generate a model for O-detector, and determine the F score using the validation dataset. According to the F scores, determine the optimal C value. Apply the corresponding model to the testing dataset and report the following metrics testing accuracy F score ROC curve the area under the curve 2. Logistic regression with L1 penalty term with C= 10, n--10:1:10. For each C, generate a model for O-detector, and determine the F score using the validation dataset. According to the F scores, determine the optimal C value. Apply the corresponding model to the testing dataset and report the following metrics testing accuracy F score ROC curve the area under the curve