JD의 블로그

1. 케라스의 함수형 API 케라스의 Sequential 모델은 네트워크 입력과 출력이 하나라고 가정한다. 따라서 이 모델은 층을 차례대로 쌓아 구성하는데 많은 경우 이런 과정이 적절하다. 하지만 이런 가정이 맞지 않는 경우도 많다. 일부 네트워크는 개별 입력이 여러 개 필요하거나 출력이 여러 개 필요하다. 예를 들어, 다양한 입력 소스에서 전달된 데이터를 다른 종류의 신경망 층을 사용하여 처리하고 합친다. 중고 의류의 시장 가격을 예측할 때 의류 브랜드, 연도 , 제품 사진 등을 입력으로 사영하여 원-핫 인코딩으로 바꾸고 완전 연결 네트워크를 사용하여 가격을 예측할 수 있다. 이럴 경우 텍스트는 RNN이나 1D 컨브넷을 사진은 2D 컨브넷을 이용할 수 있다. 가장 간단히 세 모델을 동시에 사용하는 방법..

1. Diagnonal Matrix Most of the numbers are zeros and only the elements in their diagonal are non zeros M[i, j] = 0 if i!=j we want to store only non-zero elements A[5] = {3, 7 , 4 , 9, 6} [5x5] diagonal matrix M[i, j] if(i==j) a[i-1]; void set(int A[], int i, int j, int x) x is the element that i want to store { if(i==j) A[i-1] = x; } int get(int A[], int i, int j) { if(i==j) return A[i-1]; els..

1. Time complexity : basically depends on the procedure that you are adopting. n : some number of elements O(n), bigO to measure time consuming to take a task. For finding the time complexity either you can measure the time based on the work that you are doing, means according to your procedure, if you're clear with your procedure, you can know the time, or else from the code, program code also ..

1. What is ADT Data type 1) Representation of data (How you are storing the data) 2) operations on data (What are the operations that you allow on the data) Abstract : hiding internal details. (we need not know internal details - how these operations are performed.) Without knowing them, we can use them. When the Object Oriented Programming languages being started to use in Software Development,..

1. Physical : for storing the data [1] Array A collection of continous memory locations, all these locations are side by side. This array will have fixed size, once it is created of some size, then that size cannot be increased or decreased. The size of array is fixed. An array can be created either inside stack or heap. When should I use it? When you are sure, what is the maximum number of elem..

1. About Main Memory Memory is divided into smaller addressable units that are called as bytes. Every bytes have its own address. The entire memory is not used as a single unit, but it is divided into manageble pices, that are called as segment. 2. How a Program use memory Main mamory is divided into three sections : code section, Stack, and Heap. If I have a program file on the hard disk and if..