일상생활 썸네일형 리스트형 Leave No Context Behind： Efficient Infinite Context Transformers with Infini attention https://www.youtube.com/watch?v=r_UBBfTPcF0https://arxiv.org/abs/2404.07143 Leave No Context Behind: Efficient Infinite Context Transformers with Infini-attentionThis work introduces an efficient method to scale Transformer-based Large Language Models (LLMs) to infinitely long inputs with bounded memory and computation. A key component in our proposed approach is a new attention technique dubbed.. 더보기 Apple Shocks Again： Introducing OpenELM Open Source AI Model That Changes Everything! https://www.youtube.com/watch?v=huH0fKmw0H0 Rmsnorm 사용, swiGLU 사용, Grouped query attention LR 공개 optimizer 공개 weight decay 공개 rmsnorm 사용 - 정확하지만 느림https://arxiv.org/abs/2404.14619 OpenELM: An Efficient Language Model Family with Open-source Training and Inference FrameworkThe reproducibility and transparency of large language models are crucial for advancing open research, ensuring the trustwort.. 더보기 Beyond A*: Better Planning with Transformers via Search Dynamics Bootstrapping (Searchformer) https://www.youtube.com/watch?v=PW4JiJ-WaY4https://arxiv.org/abs/2402.14083 Beyond A*: Better Planning with Transformers via Search Dynamics BootstrappingWhile Transformers have enabled tremendous progress in various application settings, such architectures still lag behind traditional symbolic planners for solving complex decision making tasks. In this work, we demonstrate how to train Transfor.. 더보기 Anthropic's Claude 3는 자의식이 있는가? https://www.youtube.com/watch?v=GBOE9fVVVSM& 결론: 없다실험결과를 보면 그냥 소설 잘쓰는 AI 1이다. 끝 더보기 4월 4주차 AI 모음 참조 https://www.youtube.com/watch?v=ZcoOW8nqVP8 https://www.youtube.com/watch?v=Crd1SR9ibRI 허깅 페이스가 해킹된 사건 사건: 인기 있는 AI 모델 플랫폼인 허깅 페이스에서 보안 침해가 발생하여 인프라 취약점이 드러났습니다. 악용: Wiz 리서치 팀은 임의 코드를 실행할 수 있는 악의적인 모델을 업로드하여 보안 허점을 노출시켰습니다. 허깅 페이스의 대응: 허깅 페이스는 새로운 형식인 "세이프 텐서(Safe Tensors)"를 도입하여 모델 로딩 시 안전하지 않은 코드를 실행할 수 없게 했습니다. 보안 조치 세이프 텐서: 이 새로운 형식은 모델이 임의 코드를 실행할 수 없도록 하여 보안을 강화합니다. 모델 스캐닝: 허깅 페이스는 잠재적으.. 더보기 Dijkstra's Hidden Prime Finding Algorithm https://www.youtube.com/watch?app=desktop&v=fwxjMKBMR7s 몰랐던걸 알았다... import heapq import time # dijkstraPrimes_heapq_2 함수 정의 def dijkstraPrimes_heapq_2(n): pool = [(4, 2)] primes = [2] for i in range(3, n): current_value, current_prime = heapq.heappop(pool) if current_value > i: heapq.heappush(pool, (i**2, i)) primes.append(i) else: while current_value 더보기 VAE - 간단 설명 (동영상 공) https://www.youtube.com/watch?v=q-n2HNan9jo 더보기 InstantID https://instantid.github.io/ InstantID There has been significant progress in personalized image synthesis with methods such as Textual Inversion, DreamBooth, and LoRA. Yet, their real-world applicability is hindered by high storage demands, lengthy fine-tuning processes, and the need for multi instantid.github.io https://github.com/InstantID/InstantID GitHub - InstantID/InstantID: InstantID : Z.. 더보기 이전 1 2 3 4 5 6 ··· 9 다음
