전체 글 썸네일형 리스트형 최신 MoE 모델 아키텍처 리뷰 https://www.reddit.com/r/LocalLLaMA/comments/1kldquv/architecture_review_of_the_new_moe_models/ From the LocalLLaMA community on RedditExplore this post and more from the LocalLLaMA communitywww.reddit.com 논의DeepSeek V3가 공개된 이후, 새로운 MoE(Mixture of Experts) 모델들이 급격히 등장하고 있습니다. 본 리뷰에서는 각 모델의 논문과 config.json, modeling_*.py 파일들을 검토한 후, 주요 정보를 아래 표에 정리하였습니다. 이에 기반한 주요 관찰 사항은 다음과 같습니다:DeepSeek은 V2에서 M.. 더보기 Sound event detection using weakly labeled dataset with stacked convolutional and recurrent neural network https://arxiv.org/abs/1710.02998 Sound event detection using weakly labeled dataset with stacked convolutional and recurrent neural networkThis paper proposes a neural network architecture and training scheme to learn the start and end time of sound events (strong labels) in an audio recording given just the list of sound events existing in the audio without time information (weak labels). Wearx.. 더보기 autoregressive decoding 더보기 Perceiver: General Perception with Iterative Attention https://arxiv.org/abs/2103.03206 Perceiver: General Perception with Iterative AttentionBiological systems perceive the world by simultaneously processing high-dimensional inputs from modalities as diverse as vision, audition, touch, proprioception, etc. The perception models used in deep learning on the other hand are designed for individualarxiv.org MLA 개념의 최초 제안 초록생물학적 시스템은 시각, 청각, 촉각, 고유 감각(p.. 더보기 AlphaEvolve: A Gemini-powered coding agent for designing advanced algorithms https://deepmind.google/discover/blog/alphaevolve-a-gemini-powered-coding-agent-for-designing-advanced-algorithms/ AlphaEvolve: A Gemini-powered coding agent for designing advanced algorithmsNew AI agent evolves algorithms for math and practical applications in computing by combining the creativity of large language models with automated evaluatorsdeepmind.google 새로운 AI 에이전트, 수학 및 컴퓨팅 실용 문제 해결을.. 더보기 Absolute Zero: Reinforced Self-play Reasoning with Zero Data https://www.arxiv.org/abs/2505.03335 Absolute Zero: Reinforced Self-play Reasoning with Zero DataReinforcement learning with verifiable rewards (RLVR) has shown promise in enhancing the reasoning capabilities of large language models by learning directly from outcome-based rewards. Recent RLVR works that operate under the zero setting avoid supervisioarxiv.org 검증 가능한 보상(Verifiable Rewards)을 활용한 .. 더보기 AST-SED: An Effective Sound Event Detection Method Based on Audio Spectrogram Transformer https://arxiv.org/abs/2303.03689 AST-SED: An Effective Sound Event Detection Method Based on Audio Spectrogram TransformerIn this paper, we propose an effective sound event detection (SED) method based on the audio spectrogram transformer (AST) model, pretrained on the large-scale AudioSet for audio tagging (AT) task, termed AST-SED. Pretrained AST models have recently shownarxiv.org 초록본 논문에서는 대.. 더보기 Symbolic Discovery of Optimization Algorithms https://arxiv.org/abs/2302.06675 Symbolic Discovery of Optimization AlgorithmsWe present a method to formulate algorithm discovery as program search, and apply it to discover optimization algorithms for deep neural network training. We leverage efficient search techniques to explore an infinite and sparse program space. To bridge tharxiv.org 초록우리는 알고리즘 발견을 프로그램 탐색(program search) 문제로 공식화하는 방법을 제.. 더보기 이전 1 2 3 4 ··· 67 다음
