opencv

Commit Graph

Author	SHA1	Message	Date
Yulv-git	15ac54d5d6	Fix some typos in modules/.	3 years ago
rogday	a2b84e9897	add assert to tf graph simplifier to address security concerns	3 years ago
OpenCV Developers	e3a55af336	dnn(test): update opencv_face_detector checks original commit: `be4a432bea`	3 years ago
Alexander Alekhin	4d927e73f1	dnn(test): update OpenVINO tests 2022.1.0	3 years ago
Alexander Alekhin	b687bc807a	dnn(test): update OpenVINO tests 2021.4.2	3 years ago
Alexander Alekhin	4c79318694	dnn: fix index access	3 years ago
rogday	93353aea70	Merge pull request #21522 from rogday:lstm Fix LSTM support in ONNX * fix LSTM and add peephole support * disable old tests * turn lambdas into functions * more hacks for c++98 * add assertions * slice fixes * backport of cuda-related fixes * address review comments	3 years ago
Egor Smirnov	375fe81311	fix slice and expand	3 years ago
Maksim Shabunin	a251474144	Update filters in ONNX tests	3 years ago
Alexander Alekhin	effce0573b	dnn: drop legacy Inference Engine NN builder API	3 years ago
Alexander Alekhin	1da48beeec	dnn(ngraph): fix output names	3 years ago
Alexander Alekhin	b57ff73086	dnn(ngraph): fix outputs handling, drop 'unconnected' logic	3 years ago
Alexander Alekhin	67978b5746	dnn(ngraph): add debuging messages	3 years ago
Alexander Alekhin	062f305d1a	dnn: don't fuse 'outputs' with OpenVINO backend	3 years ago
Alexander Alekhin	1f70d4e2a5	dnn(test): re-enable ONNX split tests for OpenVINO	3 years ago
Alexander Alekhin	aa5bc20c83	dnn(ngraph): fixup get_output_as_single_output_node() replacement patch	3 years ago
Maksim Shabunin	d1e76a34a0	3.4: Use modern OpenVINO package interface original commit: `437af37b13`	3 years ago
Alexander Alekhin	a7e6a1059c	dnn(test): fix outputs handling in ONNX conformance - ONNX output is 1 tensor per defined output instead of N tensors from outputs of "output" layer	3 years ago
Alexander Alekhin	85719a0a5d	dnn: support outputs registration under new names - fixed ONNX importer	3 years ago
Alexander Alekhin	70b0274c8e	dnn: apply hint to ignore denormals processing	3 years ago
Smirnov Egor	17b2d92a3d	add optional outputs support and fix graph links	3 years ago
Alexander Alekhin	b304730225	dnn: fix API - explicit ctors, const methods	3 years ago
Alexander Alekhin	80d9f624d0	dnn: don't use aligned load without alignment checks - weights are unaligned in dasiamprn sample (comes from numpy)	3 years ago
Alexander Alekhin	6385511e88	dnn: add checks in pooling layer implementation - to avoid out of buffer access	3 years ago
Alexander Alekhin	ed4becf007	dnn(onnx): debug dump of inputs/outputs/initializers in importer	3 years ago
Alexander Alekhin	88a18c8b6a	dnn(onnx): emit error in Shape for dynamic input	3 years ago
Alexander Alekhin	51e65db715	dnn(onnx): fix Resize inputs handling	3 years ago
Alexander Alekhin	cc02fcd889	dnn: improve debug messages, add ONNX opset version	3 years ago
Alexander Alekhin	c408157a4d	dnn: do not try to rebuilt network during setInput() - this doesn't make sense in case of multiple inputs	3 years ago
Alexander Alekhin	6e299b582a	dnn(test): decompose 'DynamicAxes' test	3 years ago
Alexander Alekhin	b1a57c4cb2	fix 3.4 links	3 years ago
rogday	0a178a687a	fix const/x in Div	3 years ago
Smirnov Egor	fec2c7e715	fix Flatten layer	3 years ago
Alexander Alekhin	60c093f086	pre: OpenCV 3.4.17 (version++)	3 years ago
Maksim Shabunin	792b7e0629	(3.4) Fixed several issues found by static analysis original commit: `a079c2eb7c`	3 years ago
Alexander Alekhin	f3ba88c87c	dnn(test): update ONNX conformance filters	3 years ago
Smirnov Egor	e97c7e042b	fix max_unpool missing attributes, add default value of keepdims in reducemean/max/sum, add support for keepdims=true in full reduction branch, add new padding type to Pad	3 years ago
rogday	4827fe86bb	Merge pull request #21088 from rogday:onnx_tests Onnx conformance tests * Add ONNX conformance tests * dnn(test): add filters for ONNX conformance tests * add filter lists for OCV backend * address review comments * move test_clip_inbounds to all_denylist * address clip issue * avoid empty lists Co-authored-by: Alexander Alekhin <alexander.a.alekhin@gmail.com>	3 years ago
rogday	1613d30544	Merge pull request #21159 from rogday:ceil_mode fix ceil_mode for Average/MaxPooling * fix ceil_mode * add a comment	3 years ago
Alexander Alekhin	bd396e1fd5	dnn(test): re-enable tests which works with OpenVINO 2021.4.x (3.4)	3 years ago
Smirnov Egor	33e97e994d	add sum of 1 input	3 years ago
Smirnov Egor	11e6848bb9	add default order to transpose	3 years ago
Smirnov Egor	829410729c	add new (Log)SoftMax simplification passes	3 years ago
Smirnov Egor	0e2a3686c0	add alpha parameter to ELU layer	3 years ago
Andrew Ryrie	ea7d4be3f8	Merge pull request #20658 from smbz:lstm_optimisation * dnn: LSTM optimisation This uses the AVX-optimised fastGEMM1T for matrix multiplications where available, instead of the standard cv::gemm. fastGEMM1T is already used by the fully-connected layer. This commit involves two minor modifications: - Use unaligned access. I don't believe this involves any performance hit in on modern CPUs (Nehalem and Bulldozer onwards) in the case where the address is actually aligned. - Allow for weight matrices where the number of columns is not a multiple of 8. I have not enabled AVX-512 as I don't have an AVX-512 CPU to test on. * Fix warning about initialisation order * Remove C++11 syntax * Fix build when AVX(2) is not available In this case the CV_TRY_X macros are defined to 0, rather than being undefined. * Minor changes as requested: - Don't check hardware support for AVX(2) when dispatch is disabled for these - Add braces * Fix out-of-bounds access in fully connected layer The old tail handling in fastGEMM1T implicitly rounded vecsize up to the next multiple of 8, and the fully connected layer implements padding up to the next multiple of 8 to cope with this. The new tail handling does not round the vecsize upwards like this but it does require that the vecsize is at least 8. To adapt to the new tail handling, the fully connected layer now rounds vecsize itself at the same time as adding the padding(which makes more sense anyway). This also means that the fully connected layer always passes a vecsize of at least 8 to fastGEMM1T, which fixes the out-of-bounds access problems. * Improve tail mask handling - Use static array for generating tail masks (as requested) - Apply tail mask to the weights as well as the input vectors to prevent spurious propagation of NaNs/Infs * Revert whitespace change * Improve readability of conditions for using AVX * dnn(lstm): minor coding style changes, replaced left aligned load	3 years ago
Smirnov Egor	05db8784ae	fix Clip, LeakyReLU, LRN, Split defaults	3 years ago
Alexander Alekhin	58b06222ff	dnn(DataLayer): fix CPU/OpenCL code paths for FP16 handling	3 years ago
Alexander Alekhin	58dc397930	dnn(test): add two_inputs test with FP32/U8 data types - remove similar test from IE scope under HAVE_INF_ENGINE	3 years ago
yuki takehara	a6277370ca	Merge pull request #21107 from take1014:remove_assert_21038 resolves #21038 * remove C assert * revert C header * fix several points in review * fix test_ds.cpp	3 years ago
Alexander Alekhin	985aa0423d	dnn(test): update InferenceEngine tests	3 years ago

1 2 3 4 5 ...

1386 Commits (ea1c970190b949d677d72e17b6c8f73cdcc11c5a)