Notice
Recent Posts
Recent Comments
Link
일 | 월 | 화 | 수 | 목 | 금 | 토 |
---|---|---|---|---|---|---|
1 | 2 | 3 | 4 | 5 | ||
6 | 7 | 8 | 9 | 10 | 11 | 12 |
13 | 14 | 15 | 16 | 17 | 18 | 19 |
20 | 21 | 22 | 23 | 24 | 25 | 26 |
27 | 28 | 29 | 30 |
Tags
- 토픽모델링
- 리뷰
- nccl 업데이트
- 불규칙적 샘플링
- Time Series
- gru-d
- nccl 설치
- pre-trained llm
- NTMs
- 의료정보
- moirai
- pytorch
- ERD
- 분산 학습
- 딥러닝
- doubleml
- causal reasoning
- 패혈증 관련 급성 호흡곤란 증후군
- netflix thumbnail
- first pg on this rank that detected no heartbeat of its watchdog.
- machine learning
- timesfm
- timellm
- causal ml
- irregularly sampled time series
- causal inference
- nccl 업그레이드
- multi gpu
- GaN
- Transformer
Archives
- Today
- Total
목록기울기 (1)
데알못정을

파이토치의 구성 torch numpy와 같은 구성 torch.autograd 자동 미분을 위한 함수 포함 torch.nn 신경망을 구축하기 위한 다양한 데이터 구조나 레이어등이 정의돼 있다.(활성함수, MSEloss 등) torch.optim SGD(확률적 경사하강법) 을 중심으로 한 파라미터 최적화 알고리즘이 구현돼 있다 torch.utils.data SGD의 반복 연산을 실핼할 때 사용하는 미니 배치용 유틸리티 함수가 포함돼 있다. 1. torch.nn , optim 모듈을 사용하지 않고 수동으로 SGD를 구현하기 import torch y = 1 + 2x_1 + 3x_2 #우리가 만들 함수 w_true = torch.Tensor([1,2,3]) # 정답 파라미터 X = torch.cat([torch..
Coding
2022. 9. 30. 17:15