https://www.reddit.com/r/LocalLLaMA/comments/188boew/multiple_4090s_instead_of_h100/ From the LocalLLaMA community on RedditExplore this post and more from the LocalLLaMA communitywww.reddit.com https://bizon-tech.com/gpu-benchmarks/NVIDIA-H100-(SXM5)-vs-NVIDIA-RTX-4090/633vs637 NVIDIA H100 (SXM5) vs. NVIDIA RTX 4090| GPU Benchmarks for AI/ML, deep learning 2024 | BIZONIn this article, we are co..

https://www.reddit.com/r/LocalLLaMA/comments/188boew/multiple_4090s_instead_of_h100/ From the LocalLLaMA community on RedditExplore this post and more from the LocalLLaMA communitywww.reddit.com 내 조건에선 4090 승!

https://www.youtube.com/watch?v=f3rFzepjlIQhttps://www.youtube.com/watch?v=gfhVAcmmMc4  https://www.youtube.com/watch?v=eIRSx-g9zi0

https://youtu.be/nb_Ido_Gl3Y?si=YTcQ91MmFSazFbDS 이거 hand tracking이랑 HOTR (human object interaction Detection), 계산대 API랑 연동해서 범죄 현황을 특정하고출구에 CCTV 하나달고 re identification 모듈을 통해 범죄자 감지하면 괜찮지 않을까? 시나리오 1. 결제를하고 바로 취소 버튼을 누르고 반품이나, 행동없이 담아서 가져간다 -> 범죄현황2. 다른 물건으로 결제버튼을 눌렀지만 더 비싼 제품을 결제한것처럼 속였다 -> 범죄현황3. 돈통에서 돈을 더 빼갔다. -> 이건 카드결제를,, ㅠ4. 스캔하는 척만하고 물건을 옮긴다 -> 범죄현황

결제를 하려고 보니 결제 버튼이 색이 옅어서뭐 잘못누른줄알았다UI적인 개선이 필요해보였다.

https://github.com/changh95/WeeklySpatialAI/issues/5 5분간 DUSt3R + 렌더링. RTX 2080 Ti. phone_2_view.mp4 Street Gaussians 코드 릴리즈 seq_04.mp4 Moon 2024 - Expressive W..." data-og-host="github.com" data-og-source-url="https://github.com/changh95/WeeklySpatialAI/issues/5" data-og-url="https://github.com/changh95/WeeklySpatialAI/issues/5" data-og-image="https://scrap.kakaocdn.net/dn/bd2Du2/hyWKC2JAuQ/MYr..

def convert_mask_to_polygon(mask): contours = None if int(cv2.__version__.split('.')[0]) > 3: contours = cv2.findContours(mask, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_TC89_KCOS)[0] else: contours = cv2.findContours(mask, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_TC89_KCOS)[1] contours = max(contours, key=lambda arr: arr.size) if contours.shape.count(1): contours = np.squeeze(contours) if contours.size 3 ..

Occupancy as Set of Pointsmulti view camera로만 point cloud를 만드는 듯?Unique3D3d mesh generator같은데 성능이 괜찮아 보이네MeshAnything: Artist-Created Mesh Generation with Autoregressive TransformersAnything 모델은 언제나 흥미로워4M-21: An Any-to-Any Vision Model for Tens of Tasks and Modalitiesvision model의 지향점이 아닐까 A decoder-only foundation model for time-series forecasting Diffusion Models for Monocular Depth Estimatio..

CMake 파일 디버깅 main file debuging  인자 집어넣기 .vscode/settings.json{ "files.associations": { "string": "cpp", "ostream": "cpp" }, "cmake.debugConfig": { "args": [ "0", "0" ] } }