* [GPU] Enable deconv with oneDNN
remove post-op data_type into oneDNN.
Signed-off-by: hyunback <hyunback.kim@intel.com>
* Update to use data_type in conv sum post-op.
Signed-off-by: hyunback <hyunback.kim@intel.com>
* checking the network batchability (internal helper func on top of batch tracking) before doing hetero
* more general logic with respect to batch-ability of the network
* a dynamism check that I've owed from the PR-10560
* using the DO-detached mechanism for early hetero exit, also fixed this flag in the Batching plugin (although minor, as the DO is removed by HETERO)
* adding the dimension tracking logic depending on whether implicitly/expicitly the auto-batching is enabled
* changed the DetectionOutput affinity markup to go over results, also accomodate Convert, so only 2 subgraphs are made by the HETERO
* [MO] Print information about new API 2.0
Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>
* Apply feedback
Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>
* Apply feedback
Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>
* installing-openvino-yocto: fix documentation links
Point to the new Yocto docs website.
Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
* Update installing-openvino-yocto.md
* installing-openvino-yocto: add step to checkout specific branch
Request users to checkout specific branch of meta-intel where this
version of OpenVINO is available.
Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
Co-authored-by: Yuan Xu <yuan1.xu@intel.com>
Co-authored-by: Anuj Mittal <anuj.mittal@intel.com>
Co-authored-by: Yuan Xu <yuan1.xu@intel.com>
* Revised unique ID setting scheme. Previously it was using program id to distinguish the loop body networks' id.
However, it results in cl cache miss for same network loaded multiple time, because program ids are differnt.
Now revised it to use parent primitive id instead of program_id for unique id of nodes in body networks.
* Revised adding unique_id to entry points to have a temporal number as unique id
* Revert the canceld change
* Added test to check whether two networks loaded from same function creates same cl cache
* [GPU] Fix permute performance degradation
Signed-off-by: Andrew Kwangwoong Park <andrew.kwangwoong.park@intel.com>
* add description for update
Signed-off-by: Andrew Kwangwoong Park <andrew.kwangwoong.park@intel.com>
Transformation insert Transpose for MatMul's weights and
sets its transpose_b attribute to true.
If executed by MO, it helps to reduce LoadNetwork time on CPU plugin,
since ConvertMatMulToFC doesn't have to insert Transpose by itself.
Ticket: 78635
* support config key device priority
for example:
if AUTO:CPU,GPU
the priority of CPU will be higher than GPU
Signed-off-by: Hu, Yuan2 <yuan2.hu@intel.com>
* add test and fix compile and test error
Signed-off-by: Hu, Yuan2 <yuan2.hu@intel.com>
* add an info for device priority and add lost [AUTOPLUGIN] on log
Signed-off-by: Hu, Yuan2 <yuan2.hu@intel.com>
* parseMetaDevice return all DEVICE of GPU, when use AUTO:GPU
Signed-off-by: Hu, Yuan2 <yuan2.hu@intel.com>
* fix compile issue
Signed-off-by: Hu, Yuan2 <yuan2.hu@intel.com>
* modify test and add test case, fix code issue
Signed-off-by: Hu, Yuan2 <yuan2.hu@intel.com>
* fix a bug and mutli with HETERO test failed
Signed-off-by: Hu, Yuan2 <yuan2.hu@intel.com>
* fix mock test faild issue
Signed-off-by: Hu, Yuan2 <yuan2.hu@intel.com>
* fix misprint
Signed-off-by: Hu, Yuan2 <yuan2.hu@intel.com>
* Disable AUTO:MYRIAD case
MYRIAD/CoreThreadingTests.smoke_QueryNetwork/targetDevice=MULTI_config=MULTI_DEVICE_PRIORITIES:MYRIAD_
faild on windows
the error is
myriadFuncTests-0 INFO: [E:] [BSL] found 0 ioexpander device
Signed-off-by: Hu, Yuan2 <yuan2.hu@intel.com>
* use ov::device::priorities key in this PR
Signed-off-by: Hu, Yuan2 <yuan2.hu@intel.com>
* fix a logic bug in key_network_priority after enable device priority
add test case cover it
Signed-off-by: Hu, Yuan2 <yuan2.hu@intel.com>
