Memory issue in learning dlib object detectors

Question

• I am using dlib to detect an object. I am training my own detector with dlib.
• But when I give more than 1000 images, the data is not trained. The process is killed by the system. I searched on the Internet but did not find a specific solution for this.
• I want to know that there is some specific template for inputting input images or any other thing that I have to consider for training data. Are there any requirements for RAM or type of images or image size.
• Any help would be appreciated. Thanks in advance.