Fluffy Shark Studio

Tokenizer 제작하기https://velog.io/@jieun9851/Tokenizer-%EC%A0%9C%EC%9E%91%ED%95%98%EA%B8%B0 WordPiece Tokenizerhttps://kaya-dev.tistory.com/47 트랜스포머의 토큰 임베딩https://seungseop.tistory.com/37 트랜스포머 구현https://cpm0722.github.io/pytorch-implementation/transformerhttps://julie-tech.tistory.com/130 nn.Embedding과 nn.Linear의 차이https://velog.io/@wjdghcks6735/PyTorch-nn.Embedding-%EA%B3%BC-nn.Linear%EC%9D%98-%..

few-shot NER에 대한 papers with codehttps://paperswithcode.com/task/few-shot-ner 고려대학교 세미나http://dmqm.korea.ac.kr/activity/seminar/389 GPT-3를 이용한 클래스 생성https://medium.com/@yongsun.yoon/%EB%8D%B0%EC%9D%B4%ED%84%B0-%EC%97%86%EC%9D%B4-ner-%EB%AA%A8%EB%8D%B8-%ED%95%99%EC%8A%B5%ED%95%98%EA%B8%B0-90c4c24953a few-shot ner에 대한 부경대 논문https://repository.pknu.ac.kr:8443/handle/2021.oak/33482 KLUE: Korean Lang..

convolutional head input data: incompleted image Im and given mask M to extract tokens one: change the input dimension three: down sample the resolution(1/8 size) numbers of convolution channels and FC dimensions to 180 for the head, body, and reconstruction modules. reason of this module 1. it is designed for fast downsampling to reduce computational complexity and memory cost. 2. empirically f..

본 논문은 추천 시스템에서 Matrix Factorzation을 이용한 Collaborative Filtering 기술을 한 단계 더 진보시켰다. 해당 논문을 읽으면서 필기한 내용을 정리한다. 논문 링크: https://ieeexplore.ieee.org/abstract/document/6748996 An Efficient Non-Negative Matrix-Factorization-Based Approach to Collaborative Filtering for Recommender Systems Matrix-factorization (MF)-based approaches prove to be highly accurate and scalable in addressing collaborative filt..

비전 트랜스포머에 대해 필기한 내용을 포스팅 한다. 논문 링크: https://arxiv.org/abs/2010.11929

Neural Radinace Fields에 대해 공부한 내용을 포스팅 한다. 논문 링크: https://arxiv.org/abs/2003.08934