[Day14] RNN

[Day14] RNN

• RNN 맛보기

  • https://n-brogrammer.tistory.com/76

• RNN - Sequential Models

  • https://n-brogrammer.tistory.com/77

• Transformer - Sequential Models

  • https://n-brogrammer.tistory.com/78

 

 

중요

LSTM의 내용 정리

nn.LSTM

 nn.LSTM(input_size=28,hidden_size=256,num_layers=3,batch_first=True)

기본적으로 쓰이는 인자들만 예시로 넣었다. 아래는 LSTM의 인자들의 설명이다. 여기서 batch_first는 LSTM의 output이 어떤식으로 나올지 이다.

input_size: The number of expected features in the input x

hidden_size: The number of features in the hidden state h

num_layers: Number of recurrent layers. E.g., setting num_layers=2

would mean stacking two LSTMs together to form a stacked LSTM,

with the second LSTM taking in outputs of the first LSTM and

computing the final results. Default: 1

bias: If False, then the layer does not use bias weights b_ih and b_hh.

Default: True

batch_first: If True, then the input and output tensors are provided

as (batch, seq, feature). Default: False

dropout: If non-zero, introduces a Dropout layer on the outputs of each

LSTM layer except the last layer, with dropout probability equal to

dropout. Default: 0

bidirectional: If True, becomes a bidirectional LSTM. Default: False

 

LSTM의 결과와 각 벡터의 shape은 다음과 같다.

N: number of batches
L: sequence lengh
Q: input dim
K: number of layers
D: LSTM feature dimension

Y,(hn,cn) = LSTM(X)
X: [N x L x Q] - N input sequnce of length L with Q dim.
Y: [N x L x D] - N output sequnce of length L with D feature dim.
hn: [K x N x D] - K (per each layer) of N final hidden state with D feature dim.
cn: [K x N x D] - K (per each layer) of N final hidden state with D cell dim.

 

 

 

피어세션

• 백준 11725번: 트리의 부모 찾기 문제를 풀고 토론 하였다.

#방법 1
from collections import defaultdict
import sys
sys.setrecursionlimit(1000000)
def find_parent(n):
    for i in node[n]:
        if result[i] == 0:
            result[i]=n
            find_parent(i)


N=int(input())
node=defaultdict(list)
result=[0]*(N+1)
result[1]=1
for _ in range(N-1):
    a,b=map(int,sys.stdin.readline().split())
    node[a].append(b)
    node[b].append(a)
find_parent(1)
print(*result[2:],sep='\n',end='')

#방법 2
import sys
from collections import defaultdict
from collections import deque

N=int(input())
node=defaultdict(list)
result=[0]*(N+1)
for _ in range(N-1):
    a,b=map(int,sys.stdin.readline().split())
    node[a].append(b)
    node[b].append(a)

q=deque()
q.appendleft(1)
result[1]=1

while q:
    parent=q.pop()
    for n in node[parent]:
        if result[n]==0:
            result[n]=parent
            q.appendleft(n)
print(*result[2:],sep='\n',end='')

 

• git에 대해서 정리하고 같이 공부를 진행함.