{"id":113469,"date":"2024-02-08T22:02:49","date_gmt":"2024-02-08T20:02:49","guid":{"rendered":"https:\/\/prohoster.info\/blog\/novosti-interneta\/vypusk-savant-0-2-7-frejmvorka-kompyuternogo-zreniya-i-glubokogo-obucheniya"},"modified":"2024-02-08T22:02:49","modified_gmt":"2024-02-08T20:02:49","slug":"vypusk-savant-0-2-7-frejmvorka-kompyuternogo-zreniya-i-glubokogo-obucheniya","status":"publish","type":"post","link":"https:\/\/prohoster.info\/sq\/blog\/news\/vypusk-savant-0-2-7-frejmvorka-kompyuternogo-zreniya-i-glubokogo-obucheniya","title":{"rendered":"L\u00ebshimi i Savant 0.2.7, nj\u00eb korniz\u00eb p\u00ebr vizionin kompjuterik dhe m\u00ebsimin e thell\u00eb","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"
Lan\u00e7imi i versionit 0.2.7 t\u00eb kuadrit Python Savant, i cili thjeshton p\u00ebrdorimin e NVIDIA DeepStream p\u00ebr zgjidhjen e problemeve q\u00eb lidhen me m\u00ebsimin e makinerive. Ky kuad\u00ebr merr p\u00ebrsip\u00ebr t\u00eb gjith\u00eb pun\u00ebn e komplikuar me GStreamer ose FFmpeg, duke lejuar fokusimin n\u00eb nd\u00ebrtimin e pipeline-ve t\u00eb optimizuara me sintaks\u00eb deklarative (YAML) dhe funksione Python. Savant lejon krijimin e pipeline-ve q\u00eb punojn\u00eb nj\u00ebsoj si n\u00eb akselerator\u00ebt n\u00eb qendrat e t\u00eb dh\u00ebnave (NVIDIA Turing, Ampere, Hopper), ashtu edhe n\u00eb pajisjet edge (NVIDIA Jetson NX, AGX Xavier, Orin NX, AGX Orin, New Nano). Me Savant, mund t\u00eb p\u00ebrpunoni leht\u00ebsisht disa flukse videosh nj\u00ebkoh\u00ebsisht dhe t\u00eb krijoni shpejt pipeline t\u00eb gatshme p\u00ebr aplikacione q\u00eb p\u00ebrdorin NVIDIA TensorRT. Kodi i projektit shp\u00ebrndahet n\u00ebn licenc\u00ebn Apache 2.0. <\/p>\n
Shembulli i pun\u00ebs me modelin e zbulimit t\u00eb bazuar n\u00eb transformatorin RT-DETR;\n
Postprocesimi CUDA me CuPy p\u00ebr YOLOV8-Seg;\n
Shembulli i integrimit t\u00eb PyTorch CUDA n\u00eb pipeline Savant;\n
Demonstrimi i pun\u00ebs me objektet e orientuara. <\/ul>\n
<\/center> <\/p>\n
Funksionet e reja:\n
\n
Integrimi me Prometheus. Pipeline mund t\u00eb eksportoj\u00eb metrika ekzekutimi n\u00eb Prometheus dhe Grafana p\u00ebr monitorimin dhe gjurmimin e performanc\u00ebs. Zhvilluesit mund t\u00eb shpallin metrika t\u00eb personalizuara, t\u00eb eksportuara s\u00eb bashku me metrikat sistemike.\n
Modaliteti i kompilimit t\u00eb modeleve. Modulet tani mund t\u00eb kompilojn\u00eb modelet e tyre n\u00eb TensorRT pa e ekzekutuar pipeline-in.\n
Menaxheri i ngjarjeve shutdown n\u00eb PyFunc. Ky API i ri lejon p\u00ebrpunimin e sakt\u00eb t\u00eb operacioneve t\u00eb p\u00ebrfundimit t\u00eb pipeline-it, duke liruar burimet dhe duke e njoftuar sistemet e jashtme p\u00ebr p\u00ebrfundimin e pun\u00ebs.\n
Filtrimi i kadrove n\u00eb hyrje dhe dalje. N\u00eb m\u00ebnyr\u00eb default, procesi pranon t\u00eb gjith\u00eb kadrot q\u00eb p\u00ebrmbajn\u00eb t\u00eb dh\u00ebna video. Me an\u00eb t\u00eb filtrimit n\u00eb hyrje dhe dalje, zhvilluesit mund t\u00eb filtrojn\u00eb t\u00eb dh\u00ebnat p\u00ebr t\u00eb p\u00ebrjashtuar p\u00ebrpunimin e tyre.\n
Pas p\u00ebrpunimi i modelit n\u00eb GPU. Fal\u00eb funksionit t\u00eb ri, zhvilluesit mund t\u00eb aksesojn\u00eb tenzorr\u00ebt e p\u00ebrkthyer t\u00eb modelit direkt nga memoria e GPU-s\u00eb, pa i ngarkuar ato n\u00eb memorien e CPU-s\u00eb dhe t'i p\u00ebrpunojn\u00eb ato me CuPy, TorchVision ose OpenCV CUDA.\n
Funksionet e p\u00ebrfaq\u00ebsimit t\u00eb memories GPU. N\u00eb k\u00ebt\u00eb version, ne ofruam funksione p\u00ebr konvertimin e bufereve t\u00eb memories midis OpenCV GpuMat, GPU-tenzoreve t\u00eb PyTorch dhe tenzoreve CuPy.\n
API p\u00ebr qasje n\u00eb statistik\u00ebn e p\u00ebrdorimit t\u00eb radh\u00ebve t\u00eb procesit. Savant lejon shtimin e radh\u00ebve mes PyFunc p\u00ebr t\u00eb implementuar p\u00ebrpunimin paralel dhe bufrimin e p\u00ebrpunimit. API i shtuar u ofron zhvilluesve akses n\u00eb radh\u00ebt e ndiera n\u00eb proces dhe lejon k\u00ebrkimin e p\u00ebrdorimit t\u00eb tyre. <\/ul>\n<\/ul>\n
N\u00eb versionin e ardhsh\u00ebm (0.3.7) planifikohet kalimi n\u00eb DeepStream 6.4 pa zgjerimin e funksionalitetit. Ideja \u00ebsht\u00eb t\u00eb arrihet nj\u00eb l\u00ebshim krejt\u00ebsisht p\u00ebrputh\u00ebs me 0.2.7, por i bazuar n\u00eb DeepStream 6.4 dhe teknologji t\u00eb p\u00ebrmir\u00ebsuar, por pa prishje t\u00eb p\u00ebrputhshm\u00ebris\u00eb n\u00eb nivel API. \n Burimi: opennet.ru<\/a> <\/p>","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"excerpt":{"rendered":"
![\"L\u00ebshimi](\"\/wp-content\/uploads\/2024\/02\/0c14b808a6767394352a4eafde317f67.jpg\")
<\/center> <\/p>\n