12 Commits

Author	SHA1	Message	Date
Alex Bezdieniezhnykh	e9a44e368d	autoconvert tensor rt engine from onnx to specific CUDA gpu	2025-04-24 16:30:21 +03:00
Alex Bezdieniezhnykh	e798af470b	read cdn yaml config from api ... automate tensorrt model conversion in case of no existing one for user's gpu	2025-04-23 23:20:08 +03:00
Alex Bezdieniezhnykh	b21f8e320f	fix bug with annotation result gradient stops ... add tensorrt engine	2025-04-02 00:29:21 +03:00
Alex Bezdieniezhnykh	6429ad62c2	refactor external clients ... put model batch size as parameter in config	2025-03-24 00:33:41 +02:00
Alex Bezdieniezhnykh	d93da15528	fix switcher between modes in DatasetExplorer.xaml	2025-03-02 21:32:31 +02:00
Alex Bezdieniezhnykh	961d2499de	fix inference ... fix small issues	2025-02-14 09:00:04 +02:00
Alex Bezdieniezhnykh	e329e5bb67	make start faster	2025-02-12 13:49:01 +02:00
Alex Bezdieniezhnykh	43cae0d03c	make cython app exit correctly	2025-02-11 20:40:49 +02:00
Alex Bezdieniezhnykh	9973a16ada	print detection results	2025-02-10 18:02:44 +02:00
Alex Bezdieniezhnykh	0f13ba384e	complete requirements.txt list ... fix build.cmd	2025-02-10 16:39:44 +02:00
Alex Bezdieniezhnykh	c1b5b5fee2	use nms in the model itself, simplify and make postprocess faster. ... make inference in batches, fix c# handling, add overlap handling	2025-02-10 14:55:00 +02:00
Alex Bezdieniezhnykh	ba3e3b4a55	move python inference to Azaion.Inference folder	2025-02-06 10:48:03 +02:00